(0 min) How do we perform efficient inference while retaining high translation quality? Existing neural machine translation models, such as Transformer, achieve high performance, but they decode words one by one, which is inefficient. Recent non-autoregressive translation models speed up the inference, but their quality is still inferior. In this work, we propose DSLP, a highly efficient and high-performance model for machine translation. The key insight is to train a non-autoregressive Transformer with Deep Supervision and feed additional Layer-wise Predictions. We conducted extensive experiments on four translation tasks (both directions of WMT'14 EN-DE and WMT'16 EN-RO). Results show that our approach consistently improves the BLEU scores compared with respective base models. Specifically, our best variant outperforms the autoregressive model on three translation tasks, while being 14.8 times more efficient in inference.

(0 min) Coupled with the availability of large scale datasets, deep learning architectures have enabled rapid progress on the Question Answering task. However, most of those datasets are in English, and the performances of state-of-the-art multilingual models are significantly lower when evaluated on non-English data. Due to high data collection costs, it is not realistic to obtain annotated data for each language one desires to support. We propose a method to improve the Cross-lingual Question Answering performance without requiring additional annotated data, leveraging Question Generation models to produce synthetic samples in a cross-lingual fashion. We show that the proposed method allows to significantly outperform the baselines trained on English data only. We report a new state-of-the-art on four multilingual datasets: MLQA, XQuAD, SQuAD-it and PIAF (fr).

(0 min) In recent years, Question Answering systems have become more popular and widely used by users. Despite the increasing popularity of these systems, the their performance is not even sufficient for textual data and requires further research. These systems consist of several parts that one of them is the Answer Selection component. This component detects the most relevant answer from a list of candidate answers. The methods presented in previous researches have attempted to provide an independent model to undertake the answer-selection task. An independent model cannot comprehend the syntactic and semantic features of questions and answers with a small training dataset. To fill this gap, language models can be employed in implementing the answer selection part. This action enables the model to have a better understanding of the language in order to understand questions and answers better than previous works. In this research, we will present the "BAS" (BERT Answer Selection) that uses the BERT language model to comprehend language. The empirical results of applying the model on the TrecQA Raw, TrecQA Clean, and WikiQA datasets demonstrate that using a robust language model such as BERT can enhance the performance. Using a more robust classifier also enhances the effect of the language model on the answer selection component. The results demonstrate that language comprehension is an essential requirement in natural language processing tasks such as answer-selection.

(0 min) Many mispronunciation detection and diagnosis (MD&D) research approaches try to exploit both the acoustic and linguistic features as input. Yet the improvement of the performance is limited, partially due to the shortage of large amount annotated training data at the phoneme level. Phonetic embeddings, extracted from ASR models trained with huge amount of word level annotations, can serve as a good representation of the content of input speech, in a noise-robust and speaker-independent manner. These embeddings, when used as implicit phonetic supplementary information, can alleviate the data shortage of explicit phoneme annotations. We propose to utilize Acoustic, Phonetic and Linguistic (APL) embedding features jointly for building a more powerful MD\&D system. Experimental results obtained on the L2-ARCTIC database show the proposed approach outperforms the baseline by 9.93%, 10.13% and 6.17% on the detection accuracy, diagnosis error rate and the F-measure, respectively.

(0 min) In recent times, we have seen an increased use of text chat for communication on social networks and smartphones. This particularly involves the use of Hindi-English code-mixed text which contains words which are not recognized in English vocabulary. We have worked on detecting emotions in these mixed data and classify the sentences in human emotions which are angry, fear, happy or sad. We have used state of the art natural language processing models and compared their performance on the dataset comprising sentences in this mixed data. The dataset was collected and annotated from sources and then used to train the models.

(0 min) While unbiased machine learning models are essential for many applications, bias is a human-defined concept that can vary across tasks. Given only input-label pairs, algorithms may lack sufficient information to distinguish stable (causal) features from unstable (spurious) features. However, related tasks often share similar biases -- an observation we may leverage to develop stable classifiers in the transfer setting. In this work, we explicitly inform the target classifier about unstable features in the source tasks. Specifically, we derive a representation that encodes the unstable features by contrasting different data environments in the source task. We achieve robustness by clustering data of the target task according to this representation and minimizing the worst-case risk across these clusters. We evaluate our method on both text and image classifications. Empirical results demonstrate that our algorithm is able to maintain robustness on the target task, outperforming the best baseline by 22.9% in absolute accuracy across 12 transfer settings. Our code is available at https://github.com/YujiaBao/Tofu.

(0 min) We present RobeCzech, a monolingual RoBERTa language representation model trained on Czech data. RoBERTa is a robustly optimized Transformer-based pretraining approach. We show that RobeCzech considerably outperforms equally-sized multilingual and Czech-trained contextualized language representation models, surpasses current state of the art in all five evaluated NLP tasks and reaches state-of-the-art results in four of them. The RobeCzech model is released publicly at https://hdl.handle.net/11234/1-3691 and https://huggingface.co/ufal/robeczech-base.

(0 min) Aspect Sentiment Triplet Extraction (ASTE) aims to extract aspect term (aspect), sentiment and opinion term (opinion) triplets from sentences and can tell a complete story, i.e., the discussed aspect, the sentiment toward the aspect, and the cause of the sentiment. ASTE is a charming task, however, one triplet extracted by ASTE only includes one opinion of the aspect, but an aspect in a sentence may have multiple corresponding opinions and one opinion only provides part of the reason why the aspect has this sentiment, as a consequence, some triplets extracted by ASTE are hard to understand, and provide erroneous information for downstream tasks. In this paper, we introduce a new task, named Aspect Sentiment Multiple Opinions Triplet Extraction (ASMOTE). ASMOTE aims to extract aspect, sentiment and multiple opinions triplets. Specifically, one triplet extracted by ASMOTE contains all opinions about the aspect and can tell the exact reason that the aspect has the sentiment. We propose an Aspect-Guided Framework (AGF) to address this task. AGF first extracts aspects, then predicts their opinions and sentiments. Moreover, with the help of the proposed Sequence Labeling Attention(SLA), AGF improves the performance of the sentiment classification using the extracted opinions. Experimental results on multiple datasets demonstrate the effectiveness of our approach.

(0 min) Conventional fine-tuning of pre-trained language models tunes all model parameters and stores a full model copy for each downstream task, which has become increasingly infeasible as the model size grows larger. Recent parameter-efficient language model tuning (PELT) methods manage to match the performance of fine-tuning with much fewer trainable parameters and perform especially well when the training data is limited. However, different PELT methods may perform rather differently on the same task, making it nontrivial to select the most appropriate method for a specific task, especially considering the fast-growing number of new PELT methods and downstream tasks. In light of model diversity and the difficulty of model selection, we propose a unified framework, UniPELT, which incorporates different PELT methods as submodules and learns to activate the ones that best suit the current data or task setup. Remarkably, on the GLUE benchmark, UniPELT consistently achieves 1~3pt gains compared to the best individual PELT method that it incorporates and even outperforms fine-tuning under different setups. Moreover, UniPELT often surpasses the upper bound when taking the best performance of all its submodules used individually on each task, indicating that a mixture of multiple PELT methods may be inherently more effective than single methods.

(0 min) Recent researches have demonstrated that BERT shows potential in a wide range of natural language processing tasks. It is adopted as an encoder for many state-of-the-art automatic summarizing systems, which achieve excellent performance. However, so far, there is not much work done for Vietnamese. In this paper, we showcase how BERT can be implemented for extractive text summarization in Vietnamese on multi-document. We introduce a novel comparison between different multilingual and monolingual BERT models. The experiment results indicate that monolingual models produce promising results compared to other multilingual models and previous text summarizing models for Vietnamese.

(0 min) Phrase representations derived from BERT often do not exhibit complex phrasal compositionality, as the model relies instead on lexical similarity to determine semantic relatedness. In this paper, we propose a contrastive fine-tuning objective that enables BERT to produce more powerful phrase embeddings. Our approach (Phrase-BERT) relies on a dataset of diverse phrasal paraphrases, which is automatically generated using a paraphrase generation model, as well as a large-scale dataset of phrases in context mined from the Books3 corpus. Phrase-BERT outperforms baselines across a variety of phrase-level similarity tasks, while also demonstrating increased lexical diversity between nearest neighbors in the vector space. Finally, as a case study, we show that Phrase-BERT embeddings can be easily integrated with a simple autoencoder to build a phrase-based neural topic model that interprets topics as mixtures of words and phrases by performing a nearest neighbor search in the embedding space. Crowdsourced evaluations demonstrate that this phrase-based topic model produces more coherent and meaningful topics than baseline word and phrase-level topic models, further validating the utility of Phrase-BERT.

(0 min) Contemporary natural language processing (NLP) revolves around learning from latent document representations, generated either implicitly by neural language models or explicitly by methods such as doc2vec or similar. One of the key properties of the obtained representations is their dimension. Whilst the commonly adopted dimensions of 256 and 768 offer sufficient performance on many tasks, it is many times unclear whether the default dimension is the most suitable choice for the subsequent downstream learning tasks. Furthermore, representation dimensions are seldom subject to hyperparameter tuning due to computational constraints. The purpose of this paper is to demonstrate that a surprisingly simple and efficient recursive compression procedure can be sufficient to both significantly compress the initial representation, but also potentially improve its performance when considering the task of text classification. Having smaller and less noisy representations is the desired property during deployment, as orders of magnitude smaller models can significantly reduce the computational overload and with it the deployment costs. We propose CoRe, a straightforward, representation learner-agnostic framework suitable for representation compression. The CoRe's performance is showcased and studied on a collection of 17 real-life corpora from biomedical, news, social media, and literary domains. We explored CoRe's behavior when considering contextual and non-contextual document representations, different compression levels, and 9 different compression algorithms. Current results based on more than 100,000 compression experiments indicate that recursive Singular Value Decomposition offers a very good trade-off between the compression efficiency and performance, making CoRe useful in many existing, representation-dependent NLP pipelines.

(0 min) The COVID-19 pandemic is a global crisis that has been testing every society and exposing the critical role of local politics in crisis response. In the United States, there has been a strong partisan divide between the Democratic and Republican party's narratives about the pandemic which resulted in polarization of individual behaviors and divergent policy adoption across regions. As shown in this case, as well as in most major social issues, strongly polarized narrative frameworks facilitate such narratives. To understand polarization and other social chasms, it is critical to dissect these diverging narratives. Here, taking the Democratic and Republican political social media posts about the pandemic as a case study, we demonstrate that a combination of computational methods can provide useful insights into the different contexts, framing, and characters and relationships that construct their narrative frameworks which individual posts source from. Leveraging a dataset of tweets from elite politicians in the U.S., we found that the Democrats' narrative tends to be more concerned with the pandemic as well as financial and social support, while the Republicans discuss more about other political entities such as China. We then perform an automatic framing analysis to characterize the ways in which they frame their narratives, where we found that the Democrats emphasize the government's role in responding to the pandemic, and the Republicans emphasize the roles of individuals and support for small businesses. Finally, we present a semantic role analysis that uncovers the important characters and relationships in their narratives as well as how they facilitate a membership categorization process. Our findings concretely expose the gaps in the "elusive consensus" between the two parties. Our methodologies may be applied to computationally study narratives in various domains.

(0 min) Automated audio captioning (AAC) is the task of automatically generating textual descriptions for general audio signals. A captioning system has to identify various information from the input signal and express it with natural language. Existing works mainly focus on investigating new methods and try to improve their performance measured on existing datasets. Having attracted attention only recently, very few works on AAC study the performance of existing pre-trained audio and natural language processing resources. In this paper, we evaluate the performance of off-the-shelf models with a Transformer-based captioning approach. We utilize the freely available Clotho dataset to compare four different pre-trained machine listening models, four word embedding models, and their combinations in many different settings. Our evaluation suggests that YAMNet combined with BERT embeddings produces the best captions. Moreover, in general, fine-tuning pre-trained word embeddings can lead to better performance. Finally, we show that sequences of audio embeddings can be processed using a Transformer encoder to produce higher-quality captions.

(0 min) Deep generative models have been widely used in several areas of NLP, and various techniques have been proposed to augment them or address their training challenges. In this paper, we propose a simple modification to Variational Autoencoders (VAEs) by using an Isotropic Gaussian Posterior (IGP) that allows for better utilisation of their latent representation space. This model avoids the sub-optimal behavior of VAEs related to inactive dimensions in the representation space. We provide both theoretical analysis, and empirical evidence on various datasets and tasks that show IGP leads to consistent improvement on several quantitative and qualitative grounds, from downstream task performance and sample efficiency to robustness. Additionally, we give insights about the representational properties encouraged by IGP and also show that its gain generalises to image domain as well.

(0 min) Training dense passage representations via contrastive learning (CL) has been shown effective for Open-Domain Passage Retrieval (ODPR). Recent studies mainly focus on optimizing this CL framework by improving the sampling strategy or extra pretraining. Different from previous studies, this work devotes itself to investigating the influence of conflicts in the widely used CL strategy in ODPR, motivated by our observation that a passage can be organized by multiple semantically different sentences, thus modeling such a passage as a unified dense vector is not optimal. We call such conflicts Contrastive Conflicts. In this work, we propose to solve it with a representation decoupling method, by decoupling the passage representations into contextual sentence-level ones, and design specific CL strategies to mediate these conflicts. Experiments on widely used datasets including Natural Questions, Trivia QA, and SQuAD verify the effectiveness of our method, especially on the dataset where the conflicting problem is severe. Our method also presents good transferability across the datasets, which further supports our idea of mediating Contrastive Conflicts.

(0 min) Prompt tuning, which only tunes continuous prompts with a frozen language model, substantially reduces per-task storage and memory usage at training. However, in the context of NLU, prior work and our results reveal that existing methods of prompt tuning do not perform well for normal-sized pre-trained models and for hard sequence tasks, indicating lack of universality. We present a novel empirical finding that properly-optimized prompt tuning can be universally effective across a wide range of model scales and NLU tasks, where it matches the performance of fine-tuning while having only 0.1\%-3\% tuned parameters. Our method P-Tuning v2 is not a new method but a version of prefix-tuning \cite{li2021prefix} optimized and adapted for NLU. Given the universality and simplicity of P-Tuning v2, we believe it can serve as an alternative for fine-tuning and a strong baseline for future research.

(0 min) Deep NLP models have been shown to learn spurious correlations, leaving them brittle to input perturbations. Recent work has shown that counterfactual or contrastive data -- i.e. minimally perturbed inputs -- can reveal these weaknesses, and that data augmentation using counterfactuals can help ameliorate them. Proposed techniques for generating counterfactuals rely on human annotations, perturbations based on simple heuristics, and meaning representation frameworks. We focus on the task of creating counterfactuals for question answering, which presents unique challenges related to world knowledge, semantic diversity, and answerability. To address these challenges, we develop a Retrieve-Generate-Filter(RGF) technique to create counterfactual evaluation and training data with minimal human supervision. Using an open-domain QA framework and question generation model trained on original task data, we create counterfactuals that are fluent, semantically diverse, and automatically labeled. Data augmentation with RGF counterfactuals improves performance on out-of-domain and challenging evaluation sets over and above existing methods, in both the reading comprehension and open-domain QA settings. Moreover, we find that RGF data leads to significant improvements in a model's robustness to local perturbations.

(0 min) Semantic parsing is the task of producing a structured meaning representation for natural language utterances or questions. Recent research has pointed out that the commonly-used sequence-to-sequence (seq2seq) semantic parsers struggle to generalize systematically, i.e. to handle examples that require recombining known knowledge in novel settings. In this work, we show that better systematic generalization can be achieved by producing the meaning representation (MR) directly as a graph and not as a sequence. To this end we propose LAGr, the Labeling Aligned Graphs algorithm that produces semantic parses by predicting node and edge labels for a complete multi-layer input-aligned graph. The strongly-supervised LAGr algorithm requires aligned graphs as inputs, whereas weakly-supervised LAGr infers alignments for originally unaligned target graphs using an approximate MAP inference procedure. On the COGS and CFQ compositional generalization benchmarks the strongly- and weakly- supervised LAGr algorithms achieve significant improvements upon the baseline seq2seq parsers.

(0 min) Neural rationale models are popular for interpretable predictions of NLP tasks. In these, a selector extracts segments of the input text, called rationales, and passes these segments to a classifier for prediction. Since the rationale is the only information accessible to the classifier, it is plausibly defined as the explanation. Is such a characterization unconditionally correct? In this paper, we argue to the contrary, with both philosophical perspectives and empirical evidence suggesting that rationale models are, perhaps, less rational and interpretable than expected. We call for more rigorous and comprehensive evaluations of these models to ensure desired properties of interpretability are indeed achieved. The code can be found at https://github.com/yimingz89/Neural-Rationale-Analysis.

(0 min) We present a memory-based model for context-dependent semantic parsing. Previous approaches focus on enabling the decoder to copy or modify the parse from the previous utterance, assuming there is a dependency between the current and previous parses. In this work, we propose to represent contextual information using an external memory. We learn a context memory controller that manages the memory by maintaining the cumulative meaning of sequential user utterances. We evaluate our approach on three semantic parsing benchmarks. Experimental results show that our model can better process context-dependent information and demonstrates improved performance without using task-specific decoders.

(0 min) During the last fifteen years, automatic text scaling has become one of the key tools of the Text as Data community in political science. Prominent text scaling algorithms, however, rely on the assumption that latent positions can be captured just by leveraging the information about word frequencies in documents under study. We challenge this traditional view and present a new, semantically aware text scaling algorithm, SemScale, which combines recent developments in the area of computational linguistics with unsupervised graph-based clustering. We conduct an extensive quantitative analysis over a collection of speeches from the European Parliament in five different languages and from two different legislative terms, and show that a scaling approach relying on semantic document representations is often better at capturing known underlying political dimensions than the established frequency-based (i.e., symbolic) scaling method. We further validate our findings through a series of experiments focused on text preprocessing and feature selection, document representation, scaling of party manifestos, and a supervised extension of our algorithm. To catalyze further research on this new branch of text scaling methods, we release a Python implementation of SemScale with all included data sets and evaluation procedures.

(0 min) This paper presents the results and analyses stemming from the first VoicePrivacy 2020 Challenge which focuses on developing anonymization solutions for speech technology. We provide a systematic overview of the challenge design with an analysis of submitted systems and evaluation results. In particular, we describe the voice anonymization task and datasets used for system development and evaluation. Also, we present different attack models and the associated objective and subjective evaluation metrics. We introduce two anonymization baselines and provide a summary description of the anonymization systems developed by the challenge participants. We report objective and subjective evaluation results for baseline and submitted systems. In addition, we present experimental results for alternative privacy metrics and attack models developed as a part of the post-evaluation analysis. Finally, we summarize our insights and observations that will influence the design of the next VoicePrivacy challenge edition and some directions for future voice anonymization research.

(0 min) Motivated by the success of T5 (Text-To-Text Transfer Transformer) in pre-training natural language processing models, we propose a unified-modal SpeechT5 framework that explores the encoder-decoder pre-training for self-supervised speech/text representation learning. The SpeechT5 framework consists of a shared encoder-decoder network and six modal-specific (speech/text) pre/post-nets. After preprocessing the speech/text input through the pre-nets, the shared encoder-decoder network models the sequence to sequence transformation, and then the post-nets generate the output in the speech/text modality based on the decoder output. Particularly, SpeechT5 can pre-train on a large scale of unlabeled speech and text data to improve the capability of the speech and textual modeling. To align the textual and speech information into a unified semantic space, we propose a cross-modal vector quantization method with random mixing-up to bridge speech and text. Extensive evaluations on a wide variety of spoken language processing tasks, including voice conversion, automatic speech recognition, text to speech, and speaker identification, show the superiority of the proposed SpeechT5 framework.

(0 min) In medical dialogue summarization, summaries must be coherent and must capture all the medically relevant information in the dialogue. However, learning effective models for summarization require large amounts of labeled data which is especially hard to obtain. We present an algorithm to create synthetic training data with an explicit focus on capturing medically relevant information. We utilize GPT-3 as the backbone of our algorithm and scale 210 human labeled examples to yield results comparable to using 6400 human labeled examples (~30x) leveraging low-shot learning and an ensemble method. In detailed experiments, we show that this approach produces high quality training data that can further be combined with human labeled data to get summaries that are strongly preferable to those produced by models trained on human data alone both in terms of medical accuracy and coherency.

(0 min) Fine-tuning all parameters of a pre-trained model has become the mainstream approach for transfer learning. To increase its efficiency and prevent catastrophic forgetting and interference, techniques like adapters and sparse fine-tuning have been developed. Adapters are modular, as they can be combined to adapt a model towards different facets of knowledge (e.g., dedicated language and/or task adapters). Sparse fine-tuning is expressive, as it controls the behavior of all model components. In this work, we introduce a new fine-tuning method with both these desirable properties. In particular, we learn sparse, real-valued masks based on a simple variant of the Lottery Ticket Hypothesis. Task-specific masks are obtained from annotated data in a source language, and language-specific masks from masked language modeling in a target language. Both these masks can then be composed with the pre-trained model. Unlike adapter-based fine-tuning, this method neither increases the number of parameters at inference time nor alters the original model architecture. Most importantly, it outperforms adapters in zero-shot cross-lingual transfer by a large margin in a series of multilingual benchmarks, including Universal Dependencies, MasakhaNER, and AmericasNLI. Based on an in-depth analysis, we additionally find that sparsity is crucial to prevent both 1) interference between the fine-tunings to be composed and 2) overfitting. We release the code and models at https://github.com/cambridgeltl/composable-sft.

(0 min) Dense retrieval (DR) methods conduct text retrieval by first encoding texts in the embedding space and then matching them by nearest neighbor search. This requires strong locality properties from the representation space, i.e, the close allocations of each small group of relevant texts, which are hard to generalize to domains without sufficient training data. In this paper, we aim to improve the generalization ability of DR models from source training domains with rich supervision signals to target domains without any relevant labels, in the zero-shot setting. To achieve that, we propose Momentum adversarial Domain Invariant Representation learning (MoDIR), which introduces a momentum method in the DR training process to train a domain classifier distinguishing source versus target, and then adversarially updates the DR encoder to learn domain invariant representations. Our experiments show that MoDIR robustly outperforms its baselines on 10+ ranking datasets from the BEIR benchmark in the zero-shot setup, with more than 10% relative gains on datasets with enough sensitivity for DR models' evaluation. Source code of this paper will be released.

(0 min) Knowledgeable FAQ chatbots are a valuable resource to any organization. While powerful and efficient retrieval-based models exist for English, it is rarely the case for other languages for which the same amount of training data is not available. In this paper, we propose a novel pre-training procedure to adapt ConveRT, an English conversational retriever model, to other languages with less training data available. We apply it for the first time to the task of Dutch FAQ answering related to the COVID-19 vaccine. We show it performs better than an open-source alternative in both a low-data regime and a high-data regime.

(0 min) Current open-domain conversational models can easily be made to talk in inadequate ways. Online learning from conversational feedback given by the conversation partner is a promising avenue for a model to improve and adapt, so as to generate fewer of these safety failures. However, current state-of-the-art models tend to react to feedback with defensive or oblivious responses. This makes for an unpleasant experience and may discourage conversation partners from giving feedback in the future. This work proposes SaFeRDialogues, a task and dataset of graceful responses to conversational feedback about safety failures. We collect a dataset of 10k dialogues demonstrating safety failures, feedback signaling them, and a response acknowledging the feedback. We show how fine-tuning on this dataset results in conversations that human raters deem considerably more likely to lead to a civil conversation, without sacrificing engagingness or general conversational ability.

(0 min) The goal of Word Sense Disambiguation (WSD) is to identify the sense of a polysemous word in a specific context. Deep-learning techniques using BERT have achieved very promising results in the field and different methods have been proposed to integrate structured knowledge to enhance performance. At the same time, an increasing number of data augmentation techniques have been proven to be useful for NLP tasks. Building upon previous works leveraging BERT and WordNet knowledge, we explore different data augmentation techniques on context-gloss pairs to improve the performance of WSD. In our experiment, we show that both sentence-level and word-level augmentation methods are effective strategies for WSD. Also, we find out that performance can be improved by adding hypernyms' glosses obtained from a lexical knowledge base. We compare and analyze different context-gloss augmentation techniques, and the results show that applying back translation on gloss performs the best.

(0 min) Neural topic models (NTMs) apply deep neural networks to topic modelling. Despite their success, NTMs generally ignore two important aspects: (1) only document-level word count information is utilized for the training, while more fine-grained sentence-level information is ignored, and (2) external semantic knowledge regarding documents, sentences and words are not exploited for the training. To address these issues, we propose a variational autoencoder (VAE) NTM model that jointly reconstructs the sentence and document word counts using combinations of bag-of-words (BoW) topical embeddings and pre-trained semantic embeddings. The pre-trained embeddings are first transformed into a common latent topical space to align their semantics with the BoW embeddings. Our model also features hierarchical KL divergence to leverage embeddings of each document to regularize those of their sentences, thereby paying more attention to semantically relevant sentences. Both quantitative and qualitative experiments have shown the efficacy of our model in 1) lowering the reconstruction errors at both the sentence and document levels, and 2) discovering more coherent topics from real-world datasets.

(0 min) Although large-scale pre-trained neural models have shown impressive performances in a variety of tasks, their ability to generate coherent text that appropriately models discourse phenomena is harder to evaluate and less understood. Given the claims of improved text generation quality across various systems, we consider the coherence evaluation of machine generated text to be one of the principal applications of coherence models that needs to be investigated. We explore training data and self-supervision objectives that result in a model that generalizes well across tasks and can be used off-the-shelf to perform such evaluations. Prior work in neural coherence modeling has primarily focused on devising new architectures, and trained the model to distinguish coherent and incoherent text through pairwise self-supervision on the permuted documents task. We instead use a basic model architecture and show significant improvements over state of the art within the same training regime. We then design a harder self-supervision objective by increasing the ratio of negative samples within a contrastive learning setup, and enhance the model further through automatic hard negative mining coupled with a large global negative queue encoded by a momentum encoder. We show empirically that increasing the density of negative samples improves the basic model, and using a global negative queue further improves and stabilizes the model while training with hard negative samples. We evaluate the coherence model on task-independent test sets that resemble real-world use cases and show significant improvements in coherence evaluations of downstream applications.

(0 min) Plug-and-play functionality allows deep learning models to adapt well to different tasks without requiring any parameters modified. Recently, prefix-tuning was shown to be a plug-and-play method on various text generation tasks by simply inserting corresponding continuous vectors into the inputs. However, sequence labeling tasks invalidate existing plug-and-play methods since different label sets demand changes to the architecture of the model classifier. In this work, we propose the use of label word prediction instead of classification to totally reuse the architecture of pre-trained models for sequence labeling tasks. Specifically, for each task, a label word set is first constructed by selecting a high-frequency word for each class respectively, and then, task-specific vectors are inserted into the inputs and optimized to manipulate the model predictions towards the corresponding label words. As a result, by simply switching the plugin vectors on the input, a frozen pre-trained language model is allowed to perform different tasks. Experimental results on three sequence labeling tasks show that the performance of the proposed method can achieve comparable performance with standard fine-tuning with only 0.1\% task-specific parameters. In addition, our method is up to 70 times faster than non-plug-and-play methods while switching different tasks under the resource-constrained scenario.

(0 min) Opinion summarization focuses on generating summaries that reflect popular opinions of multiple reviews for a single entity (e.g., a hotel or a product.) While generated summaries offer general and concise information about a particular entity, the information may be insufficient to help the user compare multiple entities. Thus, the user may still struggle with the question "Which one should I pick?" In this paper, we propose a {\em comparative opinion summarization} task, which is to generate two contrastive summaries and one common summary from two given sets of reviews from different entities. We develop a comparative summarization framework CoCoSum, which consists of two few-shot summarization models that are jointly used to generate contrastive and common summaries. Experimental results on a newly created benchmark CoCoTrip show that CoCoSum can produce high-quality contrastive and common summaries than state-of-the-art opinion summarization models.

(0 min) As the volume of long-form spoken-word content such as podcasts explodes, many platforms desire to present short, meaningful, and logically coherent segments extracted from the full content. Such segments can be consumed by users to sample content before diving in, as well as used by the platform to promote and recommend content. However, little published work is focused on the segmentation of spoken-word content, where the errors (noise) in transcripts generated by automatic speech recognition (ASR) services poses many challenges. Here we build a novel dataset of complete transcriptions of over 400 podcast episodes, in which we label the position of introductions in each episode. These introductions contain information about the episodes' topics, hosts, and guests, providing a valuable summary of the episode content, as it is created by the authors. We further augment our dataset with word substitutions to increase the amount of available training data. We train three Transformer models based on the pre-trained BERT and different augmentation strategies, which achieve significantly better performance compared with a static embedding model, showing that it is possible to capture generalized, larger-scale structural information from noisy, loosely-organized speech data. This is further demonstrated through an analysis of the models' inner architecture. Our methods and dataset can be used to facilitate future work on the structure-based segmentation of spoken-word content.

(0 min) Since the beginning of coronavirus, the disease has spread worldwide and drastically changed many aspects of the human's lifestyle. Twitter as a powerful tool can help researchers measure public health in response to COVID-19. According to the high volume of data production on social networks, automated text mining approaches can help search, read and summarize helpful information. This paper preprocessed the existing medical dataset regarding COVID-19 named CORD-19 and annotated the dataset for supervised classification tasks. At this time of the COVID-19 pandemic, we made a preprocessed dataset for the research community. This may contribute towards finding new solutions for some social interventions that COVID-19 has made. The preprocessed version of the mentioned dataset is publicly available through Github.

(0 min) Recent work (e.g. LAMA (Petroni et al., 2019)) has found that the quality of the factual information extracted from Large Language Models (LLMs) depends on the prompts used to query them. This inconsistency is problematic because different users will query LLMs for the same information using different wording, but should receive the same, accurate responses regardless. In this work we aim to address this shortcoming by introducing P-Adapters: lightweight models that sit between the embedding layer and first attention layer of LLMs. They take LLM embeddings as input and output continuous prompts that are used to query the LLM. Additionally, we investigate Mixture of Experts (MoE) models that learn a set of continuous prompts ("experts") and select one to query the LLM. They require a separate classifier trained on human-annotated data to map natural language prompts to the continuous ones. P-Adapters perform comparably to the more complex MoE models in extracting factual information from BERT and RoBERTa while eliminating the need for additional annotations. P-Adapters show between 12-26% absolute improvement in precision and 36-50% absolute improvement in consistency over a baseline of only using natural language queries. Finally, we investigate what makes a P-adapter successful and conclude that access to the LLM's embeddings of the original natural language prompt, particularly the subject of the entity pair being asked about, is a significant factor.

(0 min) Terms are linguistic signifiers of domain-specific concepts. Automated recognition of multi-word terms (MWT) in free text is a sequence labelling problem, which is commonly addressed using supervised machine learning methods. Their need for manual annotation of training data makes it difficult to port such methods across domains. FlexiTerm, on the other hand, is a fully unsupervised method for MWT recognition from domain-specific corpora. Originally implemented in Java as a proof of concept, it did not scale well, thus offering little practical value in the context of big data. In this paper, we describe its re-implementation in Python and compare the performance of these two implementations. The results demonstrated major improvements in terms of efficiency, which allow FlexiTerm to transition from the proof of concept to the production-grade application.

(0 min) Existing approaches to lifelong language learning rely on plenty of labeled data for learning a new task, which is hard to obtain in most real scenarios. Considering that humans can continually learn new tasks from a handful of examples, we expect the models also to be able to generalize well on new few-shot tasks without forgetting the previous ones. In this work, we define this more challenging yet practical problem as Lifelong Few-shot Language Learning (LFLL) and propose a unified framework for it based on prompt tuning of T5. Our framework called LFPT5 takes full advantage of PT's strong few-shot learning ability, and simultaneously trains the model as a task solver and a data generator. Before learning a new domain of the same task type, LFPT5 generates pseudo (labeled) samples of previously learned domains, and later gets trained on those samples to alleviate forgetting of previous knowledge as it learns the new domain. In addition, a KL divergence loss is minimized to achieve label consistency between the previous and the current model. While adapting to a new task type, LFPT5 includes and tunes additional prompt embeddings for the new task. With extensive experiments, we demonstrate that LFPT5 can be applied to various different types of tasks and significantly outperform previous methods in different LFLL settings.

(0 min) Event extraction is typically modeled as a multi-class classification problem where both event types and argument roles are treated as atomic symbols. These approaches are usually limited to a set of pre-defined types. We propose a novel event extraction framework that takes event types and argument roles as natural language queries to extract candidate triggers and arguments from the input text. With the rich semantics in the queries, our framework benefits from the attention mechanisms to better capture the semantic correlation between the event types or argument roles and the input text. Furthermore, the query-and-extract formulation allows our approach to leverage all available event annotations from various ontologies as a unified model. Experiments on two public benchmarks, ACE and ERE, demonstrate that our approach achieves state-of-the-art performance on each dataset and significantly outperforms existing methods on zero-shot event extraction. We will make all the programs publicly available once the paper is accepted.

(0 min) Pre-trained language models (PLMs), such as BERT and GPT, have revolutionized the field of NLP, not only in the general domain but also in the biomedical domain. Most prior efforts in building biomedical PLMs have resorted simply to domain adaptation and focused mainly on English. In this work we introduce eHealth, a biomedical PLM in Chinese built with a new pre-training framework. This new framework trains eHealth as a discriminator through both token-level and sequence-level discrimination. The former is to detect input tokens corrupted by a generator and select their original signals from plausible candidates, while the latter is to further distinguish corruptions of a same original sequence from those of the others. As such, eHealth can learn language semantics at both the token and sequence levels. Extensive experiments on 11 Chinese biomedical language understanding tasks of various forms verify the effectiveness and superiority of our approach. The pre-trained model is available to the public at \url{https://github.com/PaddlePaddle/Research/tree/master/KG/eHealth} and the code will also be released later.

(0 min) Aspect category sentiment analysis has attracted increasing research attention. The dominant methods make use of pre-trained language models by learning effective aspect category-specific representations, and adding specific output layers to its pre-trained representation. We consider a more direct way of making use of pre-trained language models, by casting the ACSA tasks into natural language generation tasks, using natural language sentences to represent the output. Our method allows more direct use of pre-trained knowledge in seq2seq language models by directly following the task setting during pre-training. Experiments on several benchmarks show that our method gives the best reported results, having large advantages in few-shot and zero-shot settings.

(0 min) Recent methods for embodied instruction following are typically trained end-to-end using imitation learning. This requires the use of expert trajectories and low-level language instructions. Such approaches assume learned hidden states will simultaneously integrate semantics from the language and vision to perform state tracking, spatial memory, exploration, and long-term planning. In contrast, we propose a modular method with structured representations that (1) builds a semantic map of the scene, and (2) performs exploration with a semantic search policy, to achieve the natural language goal. Our modular method achieves SOTA performance (24.46%) with a substantial (8.17 % absolute) gap from previous work while using less data by eschewing both expert trajectories and low-level instructions. Leveraging low-level language, however, can further increase our performance (26.49%). Our findings suggest that an explicit spatial memory and a semantic search policy can provide a stronger and more general representation for state-tracking and guidance, even in the absence of expert trajectories or low-level instructions.

(0 min) Training data for machine translation (MT) is often sourced from a multitude of large corpora that are multi-faceted in nature, e.g. containing contents from multiple domains or different levels of quality or complexity. Naturally, these facets do not occur with equal frequency, nor are they equally important for the test scenario at hand. In this work, we propose to optimize this balance jointly with MT model parameters to relieve system developers from manual schedule design. A multi-armed bandit is trained to dynamically choose between facets in a way that is most beneficial for the MT system. We evaluate it on three different multi-facet applications: balancing translationese and natural training data, or data from multiple domains or multiple language pairs. We find that bandit learning leads to competitive MT systems across tasks, and our analysis provides insights into its learned strategies and the underlying data sets.

(0 min) Even to a simple and short news headline, readers react in a multitude of ways: cognitively (e.g., inferring the writer's intent), emotionally (e.g., feeling distrust), and behaviorally (e.g., sharing the news with their friends). Such reactions are instantaneous and yet complex, as they rely on factors that go beyond interpreting the factual content the news headline. Instead, understanding reactions require pragmatic understanding of the news headline, including broader background knowledge about contentious news topics as well as commonsense reasoning about people's intents and emotional reactions. We propose Misinfo Reaction Frames, a pragmatic formalism for modeling how readers might react to a news headline cognitively, emotionally, and behaviorally. We also introduce a Misinfo Reaction Frames corpus, a dataset of over 200k news headline annotated with crowdsourced reactions focusing on global crises: the Covid-19 pandemic, climate change, and cancer. Empirical results confirm that it is indeed possible to learn the prominent patterns of readers' reactions to news headlines. We also find a potentially positive use case of our model; When we present our model generated inferences to people, we find that the machine inferences can increase readers' trust in real news while decreasing their trust in misinformation. Our work demonstrates the feasibility and the importance of pragmatic inferences of news to help enhance AI-guided misinformation detection and mitigation.

(0 min) The common practice for training commonsense models has gone from-human-to-corpus-to-machine: humans author commonsense knowledge graphs in order to train commonsense models. In this work, we investigate an alternative, from-machine-to-corpus-to-machine: general language models author these commonsense knowledge graphs to train commonsense models. Our study leads to a new framework, Symbolic Knowledge Distillation. As with prior art in Knowledge Distillation (Hinton et al., 2015), our approach uses larger models to teach smaller models. A key difference is that we distill knowledge symbolically-as text-in addition to the neural model. We also distill only one aspect-the commonsense of a general language model teacher, allowing the student to be a different type, a commonsense model. Altogether, we show that careful prompt engineering and a separately trained critic model allow us to selectively distill high-quality causal commonsense from GPT-3, a general language model. Empirical results demonstrate that, for the first time, a human-authored commonsense knowledge graph is surpassed by our automatically distilled variant in all three criteria: quantity, quality, and diversity. In addition, it results in a neural commonsense model that surpasses the teacher model's commonsense capabilities despite its 100x smaller size. We apply this to the ATOMIC resource, and share our new symbolic knowledge graph and commonsense models.

(0 min) Social concepts referring to non-physical objects--such as revolution, violence, or friendship--are powerful tools to describe, index, and query the content of visual data, including ever-growing collections of art images from the Cultural Heritage (CH) field. While much progress has been made towards complete image understanding in computer vision, automatic detection of social concepts evoked by images is still a challenge. This is partly due to the well-known semantic gap problem, worsened for social concepts given their lack of unique physical features, and reliance on more unspecific features than concrete concepts. In this paper, we propose the translation of recent cognitive theories about social concept representation into a software approach to represent them as multimodal frames, by integrating multisensory data. Our method focuses on the extraction, analysis, and integration of multimodal features from visual art material tagged with the concepts of interest. We define a conceptual model and present a novel ontology for formally representing social concepts as multimodal frames. Taking the Tate Gallery's collection as an empirical basis, we experiment our method on a corpus of art images to provide a proof of concept of its potential. We discuss further directions of research, and provide all software, data sources, and results.

(0 min) Adversarial attacks and backdoor attacks are two common security threats that hang over deep learning. Both of them harness task-irrelevant features of data in their implementation. Text style is a feature that is naturally irrelevant to most NLP tasks, and thus suitable for adversarial and backdoor attacks. In this paper, we make the first attempt to conduct adversarial and backdoor attacks based on text style transfer, which is aimed at altering the style of a sentence while preserving its meaning. We design an adversarial attack method and a backdoor attack method, and conduct extensive experiments to evaluate them. Experimental results show that popular NLP models are vulnerable to both adversarial and backdoor attacks based on text style transfer -- the attack success rates can exceed 90% without much effort. It reflects the limited ability of NLP models to handle the feature of text style that has not been widely realized. In addition, the style transfer-based adversarial and backdoor attack methods show superiority to baselines in many aspects. All the code and data of this paper can be obtained at https://github.com/thunlp/StyleAttack.

(0 min) Text autoencoders are often used for unsupervised conditional text generation by applying mappings in the latent space to change attributes to the desired values. Recently, Mai et al. (2020) proposed Emb2Emb, a method to learn these mappings in the embedding space of an autoencoder. However, their method is restricted to autoencoders with a single-vector embedding, which limits how much information can be retained. We address this issue by extending their method to Bag-of-Vectors Autoencoders (BoV-AEs), which encode the text into a variable-size bag of vectors that grows with the size of the text, as in attention-based models. This allows to encode and reconstruct much longer texts than standard autoencoders. Analogous to conventional autoencoders, we propose regularization techniques that facilitate learning meaningful operations in the latent space. Finally, we adapt for a training scheme that learns to map an input bag to an output bag, including a novel loss function and neural architecture. Our experimental evaluations on unsupervised sentiment transfer and sentence summarization show that our method performs substantially better than a standard autoencoder.

(0 min) Many legal professionals think that the explosion of information about local, regional, national, and international legislation makes their practice more costly, time-consuming, and even error-prone. The two main reasons for this are that most legislation is usually unstructured, and the tremendous amount and pace with which laws are released causes information overload in their daily tasks. In the case of the legal domain, the research community agrees that a system allowing to generate automatic responses to legal questions could substantially impact many practical implications in daily activities. The degree of usefulness is such that even a semi-automatic solution could significantly help to reduce the workload to be faced. This is mainly because a Question Answering system could be able to automatically process a massive amount of legal resources to answer a question or doubt in seconds, which means that it could save resources in the form of effort, money, and time to many professionals in the legal sector. In this work, we quantitatively and qualitatively survey the solutions that currently exist to meet this challenge.

(0 min) The sparsely-activated models have achieved great success in natural language processing through large-scale parameters and relatively low computational cost, and gradually become a feasible technique for training and implementing extremely large models. Due to the limit of communication cost, activating multiple experts is hardly affordable during training and inference. Therefore, previous work usually activate just one expert at a time to alleviate additional communication cost. Such routing mechanism limits the upper bound of model performance. In this paper, we first investigate a phenomenon that increasing the number of activated experts can boost the model performance with higher sparse ratio. To increase the number of activated experts without an increase in computational cost, we propose SAM (Switch and Mixture) routing, an efficient hierarchical routing mechanism that activates multiple experts in a same device (GPU). Our methods shed light on the training of extremely large sparse models and experiments prove that our models can achieve significant performance gain with great efficiency improvement.

(0 min) Nested entities are observed in many domains due to their compositionality, which cannot be easily recognized by the widely-used sequence labeling framework. A natural solution is to treat the task as a span classification problem. To increase performance on span representation and classification, it is crucial to effectively integrate all useful information of different formats, which we refer to heterogeneous factors including tokens, labels, boundaries, and related spans. To fuse these heterogeneous factors, we propose a novel triaffine mechanism including triaffine attention and scoring, which interacts with multiple factors in both the stages of representation and classification. Experiments results show that our proposed method achieves the state-of-the-art F1 scores on four nested NER datasets: ACE2004, ACE2005, GENIA, and KBP2017.

(0 min) Despite the increasing popularity of NLP in the humanities and social sciences, advances in model performance and complexity have been accompanied by concerns about interpretability and explanatory power for sociocultural analysis. One popular model that balances complexity and legibility is Word Mover's Distance (WMD). Ostensibly adapted for its interpretability, WMD has nonetheless been used and further developed in ways which frequently discard its most interpretable aspect: namely, the word-level distances required for translating a set of words into another set of words. To address this apparent gap, we introduce WMDecompose: a model and Python library that 1) decomposes document-level distances into their constituent word-level distances, and 2) subsequently clusters words to induce thematic elements, such that useful lexical information is retained and summarized for analysis. To illustrate its potential in a social scientific context, we apply it to a longitudinal social media corpus to explore the interrelationship between conspiracy theories and conservative American discourses. Finally, because of the full WMD model's high time-complexity, we additionally suggest a method of sampling document pairs from large datasets in a reproducible way, with tight bounds that prevent extrapolation of unreliable results due to poor sampling practices.

(0 min) Neural abstractive summarization models are susceptible to generating factually inconsistent content, a phenomenon known as hallucination. This limits the usability and adoption of these systems in real-world applications. To reduce the presence of hallucination, we propose the Mixture of Factual Experts (MoFE) model, which combines multiple summarization experts that each target a specific type of error. We train our experts using reinforcement learning (RL) to minimize the error defined by two factual consistency metrics: entity overlap and dependency arc entailment. We construct MoFE by combining the experts using two ensembling strategies (weights and logits) and evaluate them on two summarization datasets (XSUM and CNN/DM). Our experiments on BART models show that the MoFE improves performance according to both entity overlap and dependency arc entailment, without a significant performance drop on standard ROUGE metrics. The performance improvement also transfers to unseen factual consistency metrics, such as question answer-based factuality evaluation metric and BERTScore precision with respect to the source document.

(0 min) Understanding of causal narratives communicated in clinical notes can help make strides towards personalized healthcare. In this work, MIMICause, we propose annotation guidelines, develop an annotated corpus and provide baseline scores to identify types and direction of causal relations between a pair of biomedical concepts in clinical notes; communicated implicitly or explicitly, identified either in a single sentence or across multiple sentences. We annotate a total of 2714 de-identified examples sampled from the 2018 n2c2 shared task dataset and train four different language model based architectures. Annotation based on our guidelines achieved a high inter-annotator agreement i.e. Fleiss' kappa score of 0.72 and our model for identification of causal relation achieved a macro F1 score of 0.56 on test data. The high inter-annotator agreement for clinical text shows the quality of our annotation guidelines while the provided baseline F1 score sets the direction for future research towards understanding narratives in clinical texts.

(0 min) Recently, sequence-to-sequence (seq-to-seq) models have been successfully applied in text-to-speech (TTS) to synthesize speech for single-language text. To synthesize speech for multiple languages usually requires multi-lingual speech from the target speaker. However, it is both laborious and expensive to collect high-quality multi-lingual TTS data for the target speakers. In this paper, we proposed to use low-quality code-switched found data from the non-target speakers to achieve cross-lingual voice cloning for the target speakers. Experiments show that our proposed method can generate high-quality code-switched speech in the target voices in terms of both naturalness and speaker consistency. More importantly, we find that our method can achieve a comparable result to the state-of-the-art (SOTA) performance in cross-lingual voice cloning.

(0 min) In this paper, we present an extensive investigation of multi-bridge, many-to-many multilingual NMT models (MB-M2M) ie., models trained on non-English language pairs in addition to English-centric language pairs. In addition to validating previous work which shows that MB-M2M models can overcome zeroshot translation problems, our analysis reveals the following results about multibridge models: (1) it is possible to extract a reasonable amount of parallel corpora between non-English languages for low-resource languages (2) with limited non-English centric data, MB-M2M models are competitive with or outperform pivot models, (3) MB-M2M models can outperform English-Any models and perform at par with Any-English models, so a single multilingual NMT system can serve all translation directions.

(0 min) In various natural language processing tasks, passage retrieval and passage re-ranking are two key procedures in finding and ranking relevant information. Since both the two procedures contribute to the final performance, it is important to jointly optimize them in order to achieve mutual improvement. In this paper, we propose a novel joint training approach for dense passage retrieval and passage re-ranking. A major contribution is that we introduce the dynamic listwise distillation, where we design a unified listwise training approach for both the retriever and the re-ranker. During the dynamic distillation, the retriever and the re-ranker can be adaptively improved according to each other's relevance information. We also propose a hybrid data augmentation strategy to construct diverse training instances for listwise training approach. Extensive experiments show the effectiveness of our approach on both MSMARCO and Natural Questions datasets. Our code is available at https://github.com/PaddlePaddle/RocketQA.

(0 min) Style transfer is the task of rewriting an input sentence into a target style while approximately preserving its content. While most prior literature assumes access to large style-labelled corpora, recent work (Riley et al. 2021) has attempted "few-shot" style transfer using only 3-10 sentences at inference for extracting the target style. In this work we consider one such low resource setting where no datasets are available: style transfer for Indian languages. We find that existing few-shot methods perform this task poorly, with a strong tendency to copy inputs verbatim. We push the state-of-the-art for few-shot style transfer with a new method modeling the stylistic difference between paraphrases. When compared to prior work using automatic and human evaluations, our model achieves 2-3x better performance and output diversity in formality transfer and code-mixing addition across five Indian languages. Moreover, our method is better able to control the amount of style transfer using an input scalar knob. We report promising qualitative results for several attribute transfer directions, including sentiment transfer, text simplification, gender neutralization and text anonymization, all without retraining the model. Finally we found model evaluation to be difficult due to the lack of evaluation datasets and metrics for Indian languages. To facilitate further research in formality transfer for Indic languages, we crowdsource annotations for 4000 sentence pairs in four languages, and use this dataset to design our automatic evaluation suite.

(0 min) We propose a machine-translation approach to automatically generate a playlist title from a set of music tracks. We take a sequence of track IDs as input and a sequence of words in a playlist title as output, adapting the sequence-to-sequence framework based on Recurrent Neural Network (RNN) and Transformer to the music data. Considering the orderless nature of music tracks in a playlist, we propose two techniques that remove the order of the input sequence. One is data augmentation by shuffling and the other is deleting the positional encoding. We also reorganize the existing music playlist datasets to generate phrase-level playlist titles. The result shows that the Transformer models generally outperform the RNN model. Also, removing the order of input sequence improves the performance further.

(0 min) We consider language modelling (LM) as a multi-label structured prediction task by re-framing training from solely predicting a single ground-truth word to ranking a set of words which could continue a given context. To avoid annotating top-$k$ ranks, we generate them using pre-trained LMs: GPT-2, BERT, and Born-Again models. This leads to a rank-based form of knowledge distillation (KD). We also develop a method using $N$-grams to create a non-probabilistic teacher which generates the ranks without the need of a pre-trained LM. We confirm the hypotheses that we can treat LMing as a ranking task and that we can do so without the use of a pre-trained LM. We show that rank-based KD generally improves perplexity (PPL), often with statistical significance, when compared to Kullback-Leibler-based KD. Surprisingly, given the simplicity of the method, $N$-grams act as competitive teachers and achieve similar performance as using either BERT or a Born-Again model teachers. GPT-2 always acts as the best teacher, though, and using it and a Transformer-XL student on Wiki-02, rank-based KD reduces a cross-entropy baseline from 65.27 to 55.94 and against a KL-based KD of 56.70.

(0 min) Toxic speech, also known as hate speech, is regarded as one of the crucial issues plaguing online social media today. Most recent work on toxic speech detection is constrained to the modality of text with no existing work on toxicity detection from spoken utterances. In this paper, we propose a new Spoken Language Processing task of detecting toxicity from spoken speech. We introduce DeToxy, the first publicly available toxicity annotated dataset for English speech, sourced from various openly available speech databases, consisting of over 2 million utterances. Finally, we also provide analysis on how a spoken speech corpus annotated for toxicity can help facilitate the development of E2E models which better capture various prosodic cues in speech, thereby boosting toxicity classification on spoken utterances.

(0 min) Pun location is to identify the punning word (usually a word or a phrase that makes the text ambiguous) in a given short text, and pun interpretation is to find out two different meanings of the punning word. Most previous studies adopt limited word senses obtained by WSD(Word Sense Disambiguation) technique or pronunciation information in isolation to address pun location. For the task of pun interpretation, related work pays attention to various WSD algorithms. In this paper, a model called DANN (Dual-Attentive Neural Network) is proposed for pun location, effectively integrates word senses and pronunciation with context information to address two kinds of pun at the same time. Furthermore, we treat pun interpretation as a classification task and construct pungloss pairs as processing data to solve this task. Experiments on the two benchmark datasets show that our proposed methods achieve new state-of-the-art results. Our source code is available in the public code repository.

(0 min) The field of conversational agents is growing fast and there is an increasing need for algorithms that enhance natural interaction. In this work we show how we achieved state of the art results in the Keyword Spotting field by adapting and tweaking the Xception algorithm, which achieved outstanding results in several computer vision tasks. We obtained about 96\% accuracy when classifying audio clips belonging to 35 different categories, beating human annotation at the most complex tasks proposed.

(0 min) Continual learning (CL), or domain expansion, recently became a popular topic for automatic speech recognition (ASR) acoustic modeling because practical systems have to be updated frequently in order to work robustly on types of speech not observed during initial training. While sequential adaptation allows tuning a system to a new domain, it may result in performance degradation on the old domains due to catastrophic forgetting. In this work we explore regularization-based CL for neural network acoustic models trained with the lattice-free maximum mutual information (LF-MMI) criterion. We simulate domain expansion by incrementally adapting the acoustic model on different public datasets that include several accents and speaking styles. We investigate two well-known CL techniques, elastic weight consolidation (EWC) and learning without forgetting (LWF), which aim to reduce forgetting by preserving model weights or network outputs. We additionally introduce a sequence-level LWF regularization, which exploits posteriors from the denominator graph of LF-MMI to further reduce forgetting. Empirical results show that the proposed sequence-level LWF can improve the best average word error rate across all domains by up to 9.4% relative compared with using regular LWF.

(0 min) While many studies have shown that linguistic information is encoded in hidden word representations, few have studied individual neurons, to show how and in which neurons it is encoded. Among these, the common approach is to use an external probe to rank neurons according to their relevance to some linguistic attribute, and to evaluate the obtained ranking using the same probe that produced it. We show two pitfalls in this methodology: 1. It confounds distinct factors: probe quality and ranking quality. We separate them and draw conclusions on each. 2. It focuses on encoded information, rather than information that is used by the model. We show that these are not the same. We compare two recent ranking methods and a simple one we introduce, and evaluate them with regard to both of these aspects.

(0 min) Supersized pre-trained language models have pushed the accuracy of various NLP tasks to a new state-of-the-art (SOTA). Rather than pursuing the reachless SOTA accuracy, most works are pursuing improvement on other dimensions such as efficiency, leading to "Pareto SOTA". Different from accuracy, the metric for efficiency varies across different studies, making them hard to be fairly compared. To that end, this work presents ELUE (Efficient Language Understanding Evaluation), a standard evaluation, and a public leaderboard for efficient NLP models. ELUE is dedicated to depicting the Pareto Front for various language understanding tasks, such that it can tell whether and how much a method achieves Pareto improvement. Along with the benchmark, we also pre-train and release a strong baseline, ElasticBERT, whose elasticity is both static and dynamic. ElasticBERT is static in that it allows reducing model layers on demand. ElasticBERT is dynamic in that it selectively executes parts of model layers conditioned on the input. We demonstrate the ElasticBERT, despite its simplicity, outperforms or performs on par with SOTA compressed and early exiting models. The ELUE benchmark is publicly available at this http URL

(0 min) In this paper, we present a pre-trained language model (PLM) based framework called RID for conversational recommender system (CRS). RID finetunes the large-scale PLMs such as DialoGPT, together with a pre-trained Relational Graph Convolutional Network (RGCN) to encode the node representations of an item-oriented knowledge graph. The former aims to generate fluent and diverse dialogue responses based on the strong language generation ability of PLMs, while the latter is to facilitate the item recommendation by learning better node embeddings on the structural knowledge base. To unify two modules of dialogue generation and item recommendation into a PLMs-based framework, we expand the generation vocabulary of PLMs to include an extra item vocabulary, and introduces a vocabulary pointer to control when to recommend target items in the generation process. Extensive experiments on the benchmark dataset ReDial show RID significantly outperforms the state-of-the-art methods on both evaluations of dialogue and recommendation.

(0 min) Sensitivity of deep-neural models to input noise is known to be a challenging problem. In NLP, model performance often deteriorates with naturally occurring noise, such as spelling errors. To mitigate this issue, models may leverage artificially noised data. However, the amount and type of generated noise has so far been determined arbitrarily. We therefore propose to model the errors statistically from grammatical-error-correction corpora. We present a thorough evaluation of several state-of-the-art NLP systems' robustness in multiple languages, with tasks including morpho-syntactic analysis, named entity recognition, neural machine translation, a subset of the GLUE benchmark and reading comprehension. We also compare two approaches to address the performance drop: a) training the NLP models with noised data generated by our framework; and b) reducing the input noise with external system for natural language correction. The code is released at https://github.com/ufal/kazitext.

(0 min) Radiology reports are the main form of communication between radiologists and other clinicians, and contain important information for patient care. However in order to use this information for research it is necessary to convert the raw text into structured data suitable for analysis. Domain specific contextual word embeddings have been shown to achieve impressive accuracy at such natural language processing tasks in medicine. In this work we pre-trained a contextual embedding BERT model using breast radiology reports and developed a classifier that incorporated the embedding with auxiliary global textual features in order to perform a section tokenization task. This model achieved a 98% accuracy at segregating free text reports into sections of information outlined in the Breast Imaging Reporting and Data System (BI-RADS) lexicon, a significant improvement over the Classic BERT model without auxiliary information. We then evaluated whether using section tokenization improved the downstream extraction of the following fields: modality/procedure, previous cancer, menopausal status, purpose of exam, breast density and background parenchymal enhancement. Using the BERT model pre-trained on breast radiology reports combined with section tokenization resulted in an overall accuracy of 95.9% in field extraction. This is a 17% improvement compared to an overall accuracy of 78.9% for field extraction for models without section tokenization and with Classic BERT embeddings. Our work shows the strength of using BERT in radiology report analysis and the advantages of section tokenization in identifying key features of patient factors recorded in breast radiology reports.

(0 min) What would it take to teach a machine to behave ethically? While broad ethical rules may seem straightforward to state ("thou shalt not kill"), applying such rules to real-world situations is far more complex. For example, while "helping a friend" is generally a good thing to do, "helping a friend spread fake news" is not. We identify four underlying challenges towards machine ethics and norms: (1) an understanding of moral precepts and social norms; (2) the ability to perceive real-world situations visually or by reading natural language descriptions; (3) commonsense reasoning to anticipate the outcome of alternative actions in different contexts; (4) most importantly, the ability to make ethical judgments given the interplay between competing values and their grounding in different contexts (e.g., the right to freedom of expression vs. preventing the spread of fake news). Our paper begins to address these questions within the deep learning paradigm. Our prototype model, Delphi, demonstrates strong promise of language-based commonsense moral reasoning, with up to 92.1% accuracy vetted by humans. This is in stark contrast to the zero-shot performance of GPT-3 of 52.3%, which suggests that massive scale alone does not endow pre-trained neural language models with human values. Thus, we present Commonsense Norm Bank, a moral textbook customized for machines, which compiles 1.7M examples of people's ethical judgments on a broad spectrum of everyday situations. In addition to the new resources and baseline performances for future research, our study provides new insights that lead to several important open research questions: differentiating between universal human values and personal values, modeling different moral frameworks, and explainable, consistent approaches to machine ethics.

(0 min) Deep neural networks are becoming more and more popular due to their revolutionary success in diverse areas, such as computer vision, natural language processing, and speech recognition. However, the decision-making processes of these models are generally not interpretable to users. In various domains, such as healthcare, finance, or law, it is critical to know the reasons behind a decision made by an artificial intelligence system. Therefore, several directions for explaining neural models have recently been explored. In this thesis, I investigate two major directions for explaining deep neural networks. The first direction consists of feature-based post-hoc explanatory methods, that is, methods that aim to explain an already trained and fixed model (post-hoc), and that provide explanations in terms of input features, such as tokens for text and superpixels for images (feature-based). The second direction consists of self-explanatory neural models that generate natural language explanations, that is, models that have a built-in module that generates explanations for the predictions of the model.

(0 min) In attempts to develop sample-efficient algorithms, researcher have explored myriad mechanisms for collecting and exploiting feature feedback, auxiliary annotations provided for training (but not test) instances that highlight salient evidence. Examples include bounding boxes around objects and salient spans in text. Despite its intuitive appeal, feature feedback has not delivered significant gains in practical problems as assessed on iid holdout sets. However, recent works on counterfactually augmented data suggest an alternative benefit of supplemental annotations: lessening sensitivity to spurious patterns and consequently delivering gains in out-of-domain evaluations. Inspired by these findings, we hypothesize that while the numerous existing methods for incorporating feature feedback have delivered negligible in-sample gains, they may nevertheless generalize better out-of-domain. In experiments addressing sentiment analysis, we show that feature feedback methods perform significantly better on various natural out-of-domain datasets even absent differences on in-domain evaluation. By contrast, on natural language inference tasks, performance remains comparable. Finally, we compare those tasks where feature feedback does (and does not) help.

(0 min) We introduce semantic form mid-tuning, an approach for transferring semantic knowledge from semantic meaning representations into transformer-based language encoders. In mid-tuning, we learn to align the text of general sentences -- not tied to any particular inference task -- and structured semantic representations of those sentences. Our approach does not require gold annotated semantic representations. Instead, it makes use of automatically generated semantic representations, such as from off-the-shelf PropBank and FrameNet semantic parsers. We show that this alignment can be learned implicitly via classification or directly via triplet loss. Our method yields language encoders that demonstrate improved predictive performance across inference, reading comprehension, textual similarity, and other semantic tasks drawn from the GLUE, SuperGLUE, and SentEval benchmarks. We evaluate our approach on three popular baseline models, where our experimental results and analysis concludes that current pre-trained language models can further benefit from structured semantic frames with the proposed mid-tuning method, as they inject additional task-agnostic knowledge to the encoder, improving the generated embeddings as well as the linguistic properties of the given model, as evident from improvements on a popular sentence embedding toolkit and a variety of probing tasks.

(0 min) Development of speech and language technology for social good (LT4SG), especially those targeted at the welfare of marginalized communities and speakers of low-resource and under-served languages, has been a prominent theme of research within NLP, Speech, and the AI communities. Researchers have mostly relied on their individual expertise, experiences or ad hoc surveys for prioritization of language technologies that provide social good to the end-users. This has been criticized by several scholars who argue that work on LT4SG must include the target linguistic communities during the design and development process. However, none of the LT4SG work and their critiques suggest principled techniques for prioritization of the technologies and methods for inclusion of the end-user during the development cycle. Drawing inspiration from the fields of Economics, Ethics, Psychology, and Participatory Design, here we chart out a set of methodologies for prioritizing LT4SG that are aligned with the end-user preferences. We then analyze several LT4SG efforts in light of the proposed methodologies and bring out their hidden assumptions and potential pitfalls. While the current study is limited to language technologies, we believe that the principles and prioritization techniques highlighted here are applicable more broadly to AI for Social Good.

(0 min) The goal of this paper is to learn strong lip reading models that can recognise speech in silent videos. Most prior works deal with the open-set visual speech recognition problem by adapting existing automatic speech recognition techniques on top of trivially pooled visual features. Instead, in this paper we focus on the unique challenges encountered in lip reading and propose tailored solutions. To that end we make the following contributions: (1) we propose an attention-based pooling mechanism to aggregate visual speech representations; (2) we use sub-word units for lip reading for the first time and show that this allows us to better model the ambiguities of the task; (3) we propose a training pipeline that balances the lip reading performance with other key factors such as data and compute efficiency. Following the above, we obtain state-of-the-art results on the challenging LRS2 and LRS3 benchmarks when training on public datasets, and even surpass models trained on large-scale industrial datasets by using an order of magnitude less data. Our best model achieves 22.6% word error rate on the LRS2 dataset, a performance unprecedented for lip reading models, significantly reducing the performance gap between lip reading and automatic speech recognition.

(0 min) Analysis of vision-and-language models has revealed their brittleness under linguistic phenomena such as paraphrasing, negation, textual entailment, and word substitutions with synonyms or antonyms. While data augmentation techniques have been designed to mitigate against these failure modes, methods that can integrate this knowledge into the training pipeline remain under-explored. In this paper, we present \textbf{SDRO}, a model-agnostic method that utilizes a set linguistic transformations in a distributed robust optimization setting, along with an ensembling technique to leverage these transformations during inference. Experiments on benchmark datasets with images (NLVR$^2$) and video (VIOLIN) demonstrate performance improvements as well as robustness to adversarial attacks. Experiments on binary VQA explore the generalizability of this method to other V\&L tasks.

(0 min) One often wants to take an existing, trained NLP model and use it on data from a new domain. While fine-tuning or few-shot learning can be used to adapt the base model, there is no one simple recipe to getting these working; moreover, one may not have access to the original model weights if it is deployed as a black box. To this end, we study how to improve a black box model's performance on a new domain given examples from the new domain by leveraging explanations of the model's behavior. Our approach first extracts a set of features combining human intuition about the task with model attributions generated by black box interpretation techniques, and then uses a simple model to calibrate or rerank the model's predictions based on the features. We experiment with our method on two tasks, extractive question answering and natural language inference, covering adaptation from several pairs of domains. The experimental results across all the domain pairs show that explanations are useful for calibrating these models. We show that the calibration features transfer to some extent between tasks and shed light on how to effectively use them.

(0 min) An interactive instruction following task has been proposed as a benchmark for learning to map natural language instructions and first-person vision into sequences of actions to interact with objects in a 3D simulated environment. We find that an existing end-to-end neural model for this task is not robust to variations of objects and language instructions. We assume that this problem is due to the high sensitiveness of neural feature extraction to small changes in vision and language inputs. To mitigate this problem, we propose a neuro-symbolic approach that performs reasoning over high-level symbolic representations that are robust to small changes in raw inputs. Our experiments on the ALFRED dataset show that our approach significantly outperforms the existing model by 18, 52, and 73 points in the success rate on the ToggleObject, PickupObject, and SliceObject subtasks in unseen environments respectively.

(0 min) Despite considerable advancements with deep neural language models (LMs), neural text generation still suffers from degeneration: the generated text is repetitive, generic, self-contradictory, and often lacks commonsense. Our analyses on sentence-level attention patterns in LMs reveal that neural degeneration may be associated with insufficient learning of task-specific characteristics by the attention mechanism. This finding motivates on-the-fly attention modulation -- a simple but effective method that enables the injection of priors into attention computation during inference. Automatic and human evaluation results on three text generation benchmarks demonstrate that attention modulation helps LMs generate text with enhanced fluency, creativity, and commonsense reasoning, in addition to significantly reduce sentence-level repetition.

(0 min) This technical report summarizes our method for the Video-And-Language Understanding Evaluation (VALUE) challenge (https://value-benchmark.github.io/challenge\_2021.html). We propose a CLIP-Enhanced method to incorporate the image-text pretrained knowledge into downstream video-text tasks. Combined with several other improved designs, our method outperforms the state-of-the-art by $2.4\%$ ($57.58$ to $60.00$) Meta-Ave score on VALUE benchmark.

(0 min) Open-Retrieval Generative Question Answering (GenQA) is proven to deliver high-quality, natural-sounding answers in English. In this paper, we present the first generalization of the GenQA approach for the multilingual environment. To this end, we present the GenTyDiQA dataset, which extends the TyDiQA evaluation data (Clark et al., 2020) with natural-sounding, well-formed answers in Arabic, Bengali, English, Japanese, and Russian. For all these languages, we show that a GenQA sequence-to-sequence-based model outperforms a state-of-the-art Answer Sentence Selection model. We also show that a multilingually-trained model competes with, and in some cases outperforms, its monolingual counterparts. Finally, we show that our system can even compete with strong baselines, even when fed with information from a variety of languages. Essentially, our system is able to answer a question in any language of our language set using information from many languages, making it the first Language-Agnostic GenQA system.

(0 min) Distantly supervised relation extraction (DS-RE) is generally framed as a multi-instance multi-label (MI-ML) task, where the optimal aggregation of information from multiple instances is of key importance. Intra-bag attention (Lin et al., 2016) is an example of a popularly used aggregation scheme for this framework. Apart from this scheme, however, there is not much to choose from in the DS-RE literature as most of the advances in this field are focused on improving the instance-encoding step rather than the instance-aggregation step. With recent works leveraging large pre-trained language models as encoders, the increased capacity of models might allow for more flexibility in the instance-aggregation step. In this work, we explore this hypothesis and come up with a novel aggregation scheme which we call Passage-Att. Under this aggregation scheme, we combine all instances mentioning an entity pair into a "passage of instances", which is summarized independently for each relation class. These summaries are used to predict the validity of a potential triple. We show that our Passage-Att with BERT as passage encoder achieves state-of-the-art performance in three different settings (monolingual DS, monolingual DS with manually-annotated test set, multilingual DS).

(0 min) Recent work finds modern natural language processing (NLP) models relying on spurious features for prediction. Mitigating such effects is thus important. Despite this need, there is no quantitative measure to evaluate or compare the effects of different forms of spurious features in NLP. We address this gap in the literature by quantifying model sensitivity to spurious features with a causal estimand, dubbed CENT, which draws on the concept of average treatment effect from the causality literature. By conducting simulations with four prominent NLP models -- TextRNN, BERT, RoBERTa and XLNet -- we rank the models against their sensitivity to artificial injections of eight spurious features. We further hypothesize and validate that models that are more sensitive to a spurious feature will be less robust against perturbations with this feature during inference. Conversely, data augmentation with this feature improves robustness to similar perturbations. We find statistically significant inverse correlations between sensitivity and robustness, providing empirical support for our hypothesis.

(0 min) This work concentrates on reducing the RTF and word error rate of a hybrid HMM-DNN. Our baseline system uses an architecture with TDNN and LSTM layers. We find this architecture particularly useful for lightly reverberated environments. However, these models tend to demand more computation than is desirable. In this work, we explore alternate architectures employing singular value decomposition (SVD) is applied to the TDNN layers to reduce the RTF, as well as to the affine transforms of every LSTM cell. We compare this approach with specifying bottleneck layers similar to those introduced by SVD before training. Additionally, we reduced the search space of the decoding graph to make it a better fit to operate in real-time applications. We report -61.57% relative reduction in RTF and almost 1% relative decrease in WER for our architecture trained on Fisher data along with reverberated versions of this dataset in order to match one of our target test distributions.

(0 min) Visually-grounded spoken language datasets can enable models to learn cross-modal correspondences with very weak supervision. However, modern audio-visual datasets contain biases that undermine the real-world performance of models trained on that data. We introduce Spoken ObjectNet, which is designed to remove some of these biases and provide a way to better evaluate how effectively models will perform in real-world scenarios. This dataset expands upon ObjectNet, which is a bias-controlled image dataset that features similar image classes to those present in ImageNet. We detail our data collection pipeline, which features several methods to improve caption quality, including automated language model checks. Lastly, we show baseline results on image retrieval and audio retrieval tasks. These results show that models trained on other datasets and then evaluated on Spoken ObjectNet tend to perform poorly due to biases in other datasets that the models have learned. We also show evidence that the performance decrease is due to the dataset controls, and not the transfer setting.

(0 min) Recursive processing is considered a hallmark of human linguistic abilities. A recent study evaluated recursive processing in recurrent neural language models (RNN-LMs) and showed that such models perform below chance level on embedded dependencies within nested constructions -- a prototypical example of recursion in natural language. Here, we study if state-of-the-art Transformer LMs do any better. We test four different Transformer LMs on two different types of nested constructions, which differ in whether the embedded (inner) dependency is short or long range. We find that Transformers achieve near-perfect performance on short-range embedded dependencies, significantly better than previous results reported for RNN-LMs and humans. However, on long-range embedded dependencies, Transformers' performance sharply drops below chance level. Remarkably, the addition of only three words to the embedded dependency caused Transformers to fall from near-perfect to below-chance performance. Taken together, our results reveal Transformers' shortcoming when it comes to recursive, structure-based, processing.

(0 min) This paper proposes a transformer over transformer framework, called Transformer$^2$, to perform neural text segmentation. It consists of two components: bottom-level sentence encoders using pre-trained transformers, and an upper-level transformer-based segmentation model based on the sentence embeddings. The bottom-level component transfers the pre-trained knowledge learned from large external corpora under both single and pair-wise supervised NLP tasks to model the sentence embeddings for the documents. Given the sentence embeddings, the upper-level transformer is trained to recover the segmentation boundaries as well as the topic labels of each sentence. Equipped with a multi-task loss and the pre-trained knowledge, Transformer$^2$ can better capture the semantic coherence within the same segments. Our experiments show that (1) Transformer$^2$ manages to surpass state-of-the-art text segmentation models in terms of a commonly-used semantic coherence measure; (2) in most cases, both single and pair-wise pre-trained knowledge contribute to the model performance; (3) bottom-level sentence encoders pre-trained on specific languages yield better performance than those pre-trained on specific domains.

(0 min) International Phonetic Alphabet (IPA) has been widely used in cross-lingual text-to-speech (TTS) to achieve cross-lingual voice cloning (CL VC). However, IPA itself has been understudied in cross-lingual TTS. In this paper, we report some empirical findings of building a cross-lingual TTS model using IPA as inputs. Experiments show that the way to process the IPA and suprasegmental sequence has a negligible impact on the CL VC performance. Furthermore, we find that using a dataset including one speaker per language to build an IPA-based TTS system would fail CL VC since the language-unique IPA and tone/stress symbols could leak the speaker information. In addition, we experiment with different combinations of speakers in the training dataset to further investigate the effect of the number of speakers on the CL VC performance.

(0 min) When directly using existing text generation datasets for controllable generation, we are facing the problem of not having the domain knowledge and thus the aspects that could be controlled are limited.A typical example is when using CNN/Daily Mail dataset for controllable text summarization, there is no guided information on the emphasis of summary sentences. A more useful text generator should leverage both the input text and control variables to guide the generation, which can only be built with deep understanding of the domain knowledge. Motivated by this vi-sion, our paper introduces a new text generation dataset, named MReD. Our new dataset consists of 7,089 meta-reviews and all its 45k meta-review sentences are manually annotated as one of the carefully defined 9 categories, including abstract, strength, decision, etc. We present experimental results on start-of-the-art summarization models, and propose methods for controlled generation on both extractive and abstractive models using our annotated data. By exploring various settings and analaysing the model behavior with respect to the control inputs, we demonstrate the challenges and values of our dataset. MReD allows us to have a better understanding of the meta-review corpora and enlarge the research room for controllable text generation.

(0 min) The audio segmentation mismatch between training data and those seen at run-time is a major problem in direct speech translation. Indeed, while systems are usually trained on manually segmented corpora, in real use cases they are often presented with continuous audio requiring automatic (and sub-optimal) segmentation. After comparing existing techniques (VAD-based, fixed-length and hybrid segmentation methods), in this paper we propose enhanced hybrid solutions to produce better results without sacrificing latency. Through experiments on different domains and language pairs, we show that our methods outperform all the other techniques, reducing by at least 30% the gap between the traditional VAD-based approach and optimal manual segmentation.

(0 min) In recent years, researchers tend to pre-train ever-larger language models to explore the upper limit of deep models. However, large language model pre-training costs intensive computational resources and most of the models are trained from scratch without reusing the existing pre-trained models, which is wasteful. In this paper, we propose bert2BERT, which can effectively transfer the knowledge of an existing smaller pre-trained model (e.g., BERT_BASE) to a large model (e.g., BERT_LARGE) through parameter initialization and significantly improve the pre-training efficiency of the large model. Specifically, we extend the previous function-preserving on Transformer-based language model, and further improve it by proposing advanced knowledge for large model's initialization. In addition, a two-stage pre-training method is proposed to further accelerate the training process. We did extensive experiments on representative PLMs (e.g., BERT and GPT) and demonstrate that (1) our method can save a significant amount of training cost compared with baselines including learning from scratch, StackBERT and MSLT; (2) our method is generic and applicable to different types of pre-trained models. In particular, bert2BERT saves about 45% and 47% computational cost of pre-training BERT_BASE and GPT_BASE by reusing the models of almost their half sizes. The source code will be publicly available upon publication.

(0 min) Since late 2019, COVID-19 has quickly emerged as the newest biomedical domain, resulting in a surge of new information. As with other emergent domains, the discussion surrounding the topic has been rapidly changing, leading to the spread of misinformation. This has created the need for a public space for users to ask questions and receive credible, scientific answers. To fulfill this need, we turn to the task of open-domain question-answering, which we can use to efficiently find answers to free-text questions from a large set of documents. In this work, we present such a system for the emergent domain of COVID-19. Despite the small data size available, we are able to successfully train the system to retrieve answers from a large-scale corpus of published COVID-19 scientific papers. Furthermore, we incorporate effective re-ranking and question-answering techniques, such as document diversity and multiple answer spans. Our open-domain question-answering system can further act as a model for the quick development of similar systems that can be adapted and modified for other developing emergent domains.

(0 min) In this paper, a marker-based, single-person optical motion capture method (DeepMoCap) is proposed using multiple spatio-temporally aligned infrared-depth sensors and retro-reflective straps and patches (reflectors). DeepMoCap explores motion capture by automatically localizing and labeling reflectors on depth images and, subsequently, on 3D space. Introducing a non-parametric representation to encode the temporal correlation among pairs of colorized depthmaps and 3D optical flow frames, a multi-stage Fully Convolutional Network (FCN) architecture is proposed to jointly learn reflector locations and their temporal dependency among sequential frames. The extracted reflector 2D locations are spatially mapped in 3D space, resulting in robust 3D optical data extraction. The subject's motion is efficiently captured by applying a template-based fitting technique on the extracted optical data. Two datasets have been created and made publicly available for evaluation purposes; one comprising multi-view depth and 3D optical flow annotated images (DMC2.5D), and a second, consisting of spatio-temporally aligned multi-view depth images along with skeleton, inertial and ground truth MoCap data (DMC3D). The FCN model outperforms its competitors on the DMC2.5D dataset using 2D Percentage of Correct Keypoints (PCK) metric, while the motion capture outcome is evaluated against RGB-D and inertial data fusion approaches on DMC3D, outperforming the next best method by 4.5% in total 3D PCK accuracy.

(0 min) We propose a light-weight, self-supervised adaptation for a visual navigation agent to generalize to unseen environment. Given an embodied agent trained in a noiseless environment, our objective is to transfer the agent to a noisy environment where actuation and odometry sensor noise is present. Our method encourages the agent to maximize the consistency between the global maps generated at different time steps in a round-trip trajectory. The proposed task is completely self-supervised, not requiring any supervision from ground-truth pose data or explicit noise model. In addition, optimization of the task objective is extremely light-weight, as training terminates within a few minutes on a commodity GPU. Our experiments show that the proposed task helps the agent to successfully transfer to new, noisy environments. The transferred agent exhibits improved localization and mapping accuracy, further leading to enhanced performance in downstream visual navigation tasks. Moreover, we demonstrate test-time adaptation with our self-supervised task to show its potential applicability in real-world deployment.

(0 min) Visual perception in autonomous driving is a crucial part of a vehicle to navigate safely and sustainably in different traffic conditions. However, in bad weather such as heavy rain and haze, the performance of visual perception is greatly affected by several degrading effects. Recently, deep learning-based perception methods have addressed multiple degrading effects to reflect real-world bad weather cases but have shown limited success due to 1) high computational costs for deployment on mobile devices and 2) poor relevance between image enhancement and visual perception in terms of the model ability. To solve these issues, we propose a task-driven image enhancement network connected to the high-level vision task, which takes in an image corrupted by bad weather as input. Specifically, we introduce a novel low memory network to reduce most of the layer connections of dense blocks for less memory and computational cost while maintaining high performance. We also introduce a new task-driven training strategy to robustly guide the high-level task model suitable for both high-quality restoration of images and highly accurate perception. Experiment results demonstrate that the proposed method improves the performance among lane and 2D object detection, and depth estimation largely under adverse weather in terms of both low memory and accuracy.

(0 min) Deep learning (DL) based computer vision (CV) models are generally considered as black boxes due to poor interpretability. This limitation impedes efficient diagnoses or predictions of system failure, thereby precluding the widespread deployment of DLCV models in safety-critical tasks such as autonomous driving. This study is motivated by the need to enhance the interpretability of DL model in autonomous driving and therefore proposes an explainable DL-based framework that generates textual descriptions of the driving environment and makes appropriate decisions based on the generated descriptions. The proposed framework imitates the learning process of human drivers by jointly modeling the visual input (images) and natural language, while using the language to induce the visual attention in the image. The results indicate strong explainability of autonomous driving decisions obtained by focusing on relevant features from visual inputs. Furthermore, the output attention maps enhance the interpretability of the model not only by providing meaningful explanation to the model behavior but also by identifying the weakness of and potential improvement directions for the model.

(2 min) Empirical science of neural scaling laws is a rapidly growing area of significant importance to the future of machine learning, particularly in the light of recent breakthroughs achieved by large-scale pre-trained models such as GPT-3, CLIP and DALL-e. Accurately predicting the neural network performance with increasing resources such as data, compute and model size provides a more comprehensive evaluation of different approaches across multiple scales, as opposed to traditional point-wise comparisons of fixed-size models on fixed-size benchmarks, and, most importantly, allows for focus on the best-scaling, and thus most promising in the future, approaches. In this work, we consider a challenging problem of few-shot learning in image classification, especially when the target data distribution in the few-shot phase is different from the source, training, data distribution, in a sense that it includes new image classes not encountered during training. Our current main goal is to investigate how the amount of pre-training data affects the few-shot generalization performance of standard image classifiers. Our key observations are that (1) such performance improvements are well-approximated by power laws (linear log-log plots) as the training set size increases, (2) this applies to both cases of target data coming from either the same or from a different domain (i.e., new classes) as the training data, and (3) few-shot performance on new classes converges at a faster rate than the standard classification performance on previously seen classes. Our findings shed new light on the relationship between scale and generalization.

(0 min) Point completion refers to completing the missing geometries of an object from incomplete observations. Main-stream methods predict the missing shapes by decoding a global feature learned from the input point cloud, which often leads to deficient results in preserving topology consistency and surface details. In this work, we present ME-PCN, a point completion network that leverages `emptiness' in 3D shape space. Given a single depth scan, previous methods often encode the occupied partial shapes while ignoring the empty regions (e.g. holes) in depth maps. In contrast, we argue that these `emptiness' clues indicate shape boundaries that can be used to improve topology representation and detail granularity on surfaces. Specifically, our ME-PCN encodes both the occupied point cloud and the neighboring `empty points'. It estimates coarse-grained but complete and reasonable surface points in the first stage, followed by a refinement stage to produce fine-grained surface details. Comprehensive experiments verify that our ME-PCN presents better qualitative and quantitative performance against the state-of-the-art. Besides, we further prove that our `emptiness' design is lightweight and easy to embed in existing methods, which shows consistent effectiveness in improving the CD and EMD scores.

(0 min) Object detection in three-dimensional (3D) space attracts much interest from academia and industry since it is an essential task in AI-driven applications such as robotics, autonomous driving, and augmented reality. As the basic format of 3D data, the point cloud can provide detailed geometric information about the objects in the original 3D space. However, due to 3D data's sparsity and unorderedness, specially designed networks and modules are needed to process this type of data. Attention mechanism has achieved impressive performance in diverse computer vision tasks; however, it is unclear how attention modules would affect the performance of 3D point cloud object detection and what sort of attention modules could fit with the inherent properties of 3D data. This work investigates the role of the attention mechanism in 3D point cloud object detection and provides insights into the potential of different attention modules. To achieve that, we comprehensively investigate classical 2D attentions, novel 3D attentions, including the latest point cloud transformers on SUN RGB-D and ScanNetV2 datasets. Based on the detailed experiments and analysis, we conclude the effects of different attention modules. This paper is expected to serve as a reference source for benefiting attention-embedded 3D point cloud object detection. The code and trained models are available at: https://github.com/ShiQiu0419/attentions_in_3D_detection.

(0 min) In order to function in unstructured environments, robots need the ability to recognize unseen objects. We take a step in this direction by tackling the problem of segmenting unseen object instances in tabletop environments. However, the type of large-scale real-world dataset required for this task typically does not exist for most robotic settings, which motivates the use of synthetic data. Our proposed method, UOIS-Net, separately leverages synthetic RGB and synthetic depth for unseen object instance segmentation. UOIS-Net is comprised of two stages: first, it operates only on depth to produce object instance center votes in 2D or 3D and assembles them into rough initial masks. Secondly, these initial masks are refined using RGB. Surprisingly, our framework is able to learn from synthetic RGB-D data where the RGB is non-photorealistic. To train our method, we introduce a large-scale synthetic dataset of random objects on tabletops. We show that our method can produce sharp and accurate segmentation masks, outperforming state-of-the-art methods on unseen object instance segmentation. We also show that our method can segment unseen objects for robot grasping.

(0 min) We describe a simple pre-training approach for point clouds. It works in three steps: 1. Mask all points occluded in a camera view; 2. Learn an encoder-decoder model to reconstruct the occluded points; 3. Use the encoder weights as initialisation for downstream point cloud tasks. We find that even when we construct a single pre-training dataset (from ModelNet40), this pre-training method improves accuracy across different datasets and encoders, on a wide range of downstream tasks. Specifically, we show that our method outperforms previous pre-training methods in object classification, and both part-based and semantic segmentation tasks. We study the pre-trained features and find that they lead to wide downstream minima, have high transformation invariance, and have activations that are highly correlated with part labels. Code and data are available at: https://github.com/hansen7/OcCo

(0 min) Deep neural networks (DNNs) have demonstrated their great potential in recent years, exceeding the per-formance of human experts in a wide range of applications. Due to their large sizes, however, compressiontechniques such as weight quantization and pruning are usually applied before they can be accommodated onthe edge. It is generally believed that quantization leads to performance degradation, and plenty of existingworks have explored quantization strategies aiming at minimum accuracy loss. In this paper, we argue thatquantization, which essentially imposes regularization on weight representations, can sometimes help toimprove accuracy. We conduct comprehensive experiments on three widely used applications: fully con-nected network (FCN) for biomedical image segmentation, convolutional neural network (CNN) for imageclassification on ImageNet, and recurrent neural network (RNN) for automatic speech recognition, and experi-mental results show that quantization can improve the accuracy by 1%, 1.95%, 4.23% on the three applicationsrespectively with 3.5x-6.4x memory reduction.

(0 min) Human visual recognition of activities or external agents involves an interplay between high-level plan recognition and low-level perception. Given that, a natural question to ask is: can low-level perception be improved by high-level plan recognition? We formulate the problem of leveraging recognized plans to generate better top-down attention maps \cite{gazzaniga2009,baluch2011} to improve the perception performance. We call these top-down attention maps specifically as plan-recognition-driven attention maps. To address this problem, we introduce the Pixel Dynamics Network. Pixel Dynamics Network serves as an observation model, which predicts next states of object points at each pixel location given observation of pixels and pixel-level action feature. This is like internally learning a pixel-level dynamics model. Pixel Dynamics Network is a kind of Convolutional Neural Network (ConvNet), with specially-designed architecture. Therefore, Pixel Dynamics Network could take the advantage of parallel computation of ConvNets, while learning the pixel-level dynamics model. We further prove the equivalence between Pixel Dynamics Network as an observation model, and the belief update in partially observable Markov decision process (POMDP) framework. We evaluate our Pixel Dynamics Network in event recognition tasks. We build an event recognition system, ER-PRN, which takes Pixel Dynamics Network as a subroutine, to recognize events based on observations augmented by plan-recognition-driven attention.

(0 min) Automatic License Plate Recognition systems aim to provide a solution for detecting, localizing, and recognizing license plate characters from vehicles appearing in video frames. However, deploying such systems in the real world requires real-time performance in low-resource environments. In our paper, we propose a two-stage detection pipeline paired with Vision API that provides real-time inference speed along with consistently accurate detection and recognition performance. We used a haar-cascade classifier as a filter on top of our backbone MobileNet SSDv2 detection model. This reduces inference time by only focusing on high confidence detections and using them for recognition. We also impose a temporal frame separation strategy to distinguish between multiple vehicle license plates in the same clip. Furthermore, there are no publicly available Bangla license plate datasets, for which we created an image dataset and a video dataset containing license plates in the wild. We trained our models on the image dataset and achieved an AP(0.5) score of 86% and tested our pipeline on the video dataset and observed reasonable detection and recognition performance (82.7% detection rate, and 60.8% OCR F1 score) with real-time processing speed (27.2 frames per second).

(0 min) Existing neural classification networks predominately adopt one-hot encoding due to its simplicity in representing categorical data. However, the one-hot representation neglects inter-category correlations, which may result in poor generalization. Herein, we observe that a pre-trained baseline network has paid attention to the target image region even though it incorrectly predicts the image, revealing which categories confuse the baseline. This observation motivates us to consider inter-category correlations. Therefore, we propose a clonal network, named ClonalNet, which learns to discriminate between confusing categories derived from the pre-trained baseline. The ClonalNet architecture can be identical or smaller than the baseline architecture. When identical, ClonalNet is a clonal version of the baseline but does not share weights. When smaller, the training process of ClonalNet resembles that of the standard knowledge distillation. The difference from knowledge distillation is that we design a focusing-picking loss to optimize ClonalNet. This novel loss enforces ClonalNet to concentrate on confusing categories and make more confident predictions on ground-truth labels with the baseline reference. Experiments show that ClonalNet significantly outperforms baseline networks and knowledge distillation.

(0 min) Sensor drift is a long-existing unpredictable problem that deteriorates the performance of gaseous substance recognition, calling for an antidrift domain adaptation algorithm. However, the prerequisite for traditional methods to achieve fine results is to have data from both nondrift distributions (source domain) and drift distributions (target domain) for domain alignment, which is usually unrealistic and unachievable in real-life scenarios. To compensate for this, in this paper, deep learning based on a target-domain-free domain adaptation convolutional neural network (TDACNN) is proposed. The main concept is that CNNs extract not only the domain-specific features of samples but also the domain-invariant features underlying both the source and target domains. Making full use of these various levels of embedding features can lead to comprehensive utilization of different levels of characteristics, thus achieving drift compensation by the extracted intermediate features between two domains. In the TDACNN, a flexible multibranch backbone with a multiclassifier structure is proposed under the guidance of bionics, which utilizes multiple embedding features comprehensively without involving target domain data during training. A classifier ensemble method based on maximum mean discrepancy (MMD) is proposed to evaluate all the classifiers jointly based on the credibility of the pseudolabel. To optimize network training, an additive angular margin softmax loss with parameter dynamic adjustment is utilized. Experiments on two drift datasets under different settings demonstrate the superiority of TDACNN compared with several state-of-the-art methods.

(0 min) The Possibilistic Fuzzy Local Information C-Means (PFLICM) method is presented as a technique to segment side-look synthetic aperture sonar (SAS) imagery into distinct regions of the sea-floor. In this work, we investigate and present the results of an automated feature selection approach for SAS image segmentation. The chosen features and resulting segmentation from the image will be assessed based on a select quantitative clustering validity criterion and the subset of the features that reach a desired threshold will be used for the segmentation process.

(0 min) We study a new family of inverse problems for recovering representations of corrupted data. We assume access to a pre-trained representation learning network R(x) that operates on clean images, like CLIP. The problem is to recover the representation of an image R(x), if we are only given a corrupted version A(x), for some known forward operator A. We propose a supervised inversion method that uses a contrastive objective to obtain excellent representations for highly corrupted images. Using a linear probe on our robust representations, we achieve a higher accuracy than end-to-end supervised baselines when classifying images with various types of distortions, including blurring, additive noise, and random pixel masking. We evaluate on a subset of ImageNet and observe that our method is robust to varying levels of distortion. Our method outperforms end-to-end baselines even with a fraction of the labeled data in a wide range of forward operators.

(0 min) Recent works on sparse neural networks have demonstrated the possibility to train a sparse subnetwork independently from scratch, to match the performance of its corresponding dense network. However, identifying such sparse subnetworks (winning tickets) either involves a costly iterative train-prune-retrain process (e.g., Lottery Ticket Hypothesis) or an over-extended training time (e.g., Dynamic Sparse Training). In this work, we draw a unique connection between sparse neural network training and the deep ensembling technique, yielding a novel ensemble learning framework called FreeTickets. Instead of starting from a dense network, FreeTickets randomly initializes a sparse subnetwork and then trains the subnetwork while dynamically adjusting its sparse mask, resulting in many diverse sparse subnetworks throughout the training process. FreeTickets is defined as the ensemble of these sparse subnetworks freely obtained during this one-pass, sparse-to-sparse training, which uses only a fraction of the computational resources required by the vanilla dense training. Moreover, despite being an ensemble of models, FreeTickets has even fewer parameters and training FLOPs compared to a single dense model: this seemingly counter-intuitive outcome is due to the high sparsity of each subnetwork. FreeTickets is observed to demonstrate a significant all-round improvement compared to standard dense baselines, in prediction accuracy, uncertainty estimation, robustness, and efficiency. FreeTickets easily outperforms the naive deep ensemble with ResNet50 on ImageNet using only a quarter of the training FLOPs required by the latter. Our results provide insights into the strength of sparse neural networks and suggest that the benefits of sparsity go way beyond the usually expected inference efficiency.

(0 min) We present a novel method to reconstruct a spectral central view and its aligned disparity map from spatio-spectrally coded light fields. Since we do not reconstruct an intermediate full light field from the coded measurement, we refer to this as principal reconstruction. The coded light fields correspond to those captured by a light field camera in the unfocused design with a spectrally coded microlens array. In this application, the spectrally coded light field camera can be interpreted as a single-shot spectral depth camera. We investigate several multi-task deep learning methods and propose a new auxiliary loss-based training strategy to enhance the reconstruction performance. The results are evaluated using a synthetic as well as a new real-world spectral light field dataset that we captured using a custom-built camera. The results are compared to state-of-the art compressed sensing reconstruction and disparity estimation. We achieve a high reconstruction quality for both synthetic and real-world coded light fields. The disparity estimation quality is on par with or even outperforms state-of-the-art disparity estimation from uncoded RGB light fields.

(0 min) We propose a method for generalizing deep learning for 3D point cloud registration on new, totally different datasets. It is based on two components, MS-SVConv and UDGE. Using Multi-Scale Sparse Voxel Convolution, MS-SVConv is a fast deep neural network that outputs the descriptors from point clouds for 3D registration between two scenes. UDGE is an algorithm for transferring deep networks on unknown datasets in a unsupervised way. The interest of the proposed method appears while using the two components, MS-SVConv and UDGE, together as a whole, which leads to state-of-the-art results on real world registration datasets such as 3DMatch, ETH and TUM. The code is publicly available at https://github.com/humanpose1/MS-SVConv .

(0 min) Mitochondria instance segmentation from electron microscopy (EM) images has seen notable progress since the introduction of deep learning methods. In this paper, we propose two advanced deep networks, named Res-UNet-R and Res-UNet-H, for 3D mitochondria instance segmentation from Rat and Human samples. Specifically, we design a simple yet effective anisotropic convolution block and deploy a multi-scale training strategy, which together boost the segmentation performance. Moreover, we enhance the generalizability of the trained models on the test set by adding a denoising operation as pre-processing. In the Large-scale 3D Mitochondria Instance Segmentation Challenge at ISBI 2021, our method ranks the 1st place. Code is available at https://github.com/Limingxing00/MitoEM2021-Challenge.

(0 min) Improving adversarial robustness of neural networks remains a major challenge. Fundamentally, training a neural network via gradient descent is a parameter estimation problem. In adaptive control, maintaining persistency of excitation (PoE) is integral to ensuring convergence of parameter estimates in dynamical systems to their true values. We show that parameter estimation with gradient descent can be modeled as a sampling of an adaptive linear time-varying continuous system. Leveraging this model, and with inspiration from Model-Reference Adaptive Control (MRAC), we prove a sufficient condition to constrain gradient descent updates to reference persistently excited trajectories converging to the true parameters. The sufficient condition is achieved when the learning rate is less than the inverse of the Lipschitz constant of the gradient of loss function. We provide an efficient technique for estimating the corresponding Lipschitz constant in practice using extreme value theory. Our experimental results in both standard and adversarial training illustrate that networks trained with the PoE-motivated learning rate schedule have similar clean accuracy but are significantly more robust to adversarial attacks than models trained using current state-of-the-art heuristics.

(0 min) Recent monocular human performance capture approaches have shown compelling dense tracking results of the full body from a single RGB camera. However, existing methods either do not estimate clothing at all or model cloth deformation with simple geometric priors instead of taking into account the underlying physical principles. This leads to noticeable artifacts in their reconstructions, e.g. baked-in wrinkles, implausible deformations that seemingly defy gravity, and intersections between cloth and body. To address these problems, we propose a person-specific, learning-based method that integrates a simulation layer into the training process to provide for the first time physics supervision in the context of weakly supervised deep monocular human performance capture. We show how integrating physics into the training process improves the learned cloth deformations, allows modeling clothing as a separate piece of geometry, and largely reduces cloth-body intersections. Relying only on weak 2D multi-view supervision during training, our approach leads to a significant improvement over current state-of-the-art methods and is thus a clear step towards realistic monocular capture of the entire deforming surface of a clothed human.

(0 min) While unbiased machine learning models are essential for many applications, bias is a human-defined concept that can vary across tasks. Given only input-label pairs, algorithms may lack sufficient information to distinguish stable (causal) features from unstable (spurious) features. However, related tasks often share similar biases -- an observation we may leverage to develop stable classifiers in the transfer setting. In this work, we explicitly inform the target classifier about unstable features in the source tasks. Specifically, we derive a representation that encodes the unstable features by contrasting different data environments in the source task. We achieve robustness by clustering data of the target task according to this representation and minimizing the worst-case risk across these clusters. We evaluate our method on both text and image classifications. Empirical results demonstrate that our algorithm is able to maintain robustness on the target task, outperforming the best baseline by 22.9% in absolute accuracy across 12 transfer settings. Our code is available at https://github.com/YujiaBao/Tofu.

(0 min) Accurate delineation of the left ventricular boundaries in late gadolinium-enhanced magnetic resonance imaging (LGE-MRI) is an essential step for scar tissue quantification and patient-specific assessment of myocardial infarction. Many deep-learning techniques have been proposed to perform automatic segmentations of the left ventricle (LV) in LGE-MRI showing segmentations as accurate as those obtained by expert cardiologists. Thus far, the existing models have been overwhelmingly developed and evaluated with LGE-MRI datasets from single clinical centers. However, in practice, LGE-MRI images vary significantly between clinical centers within and across countries, in particular due to differences in the MRI scanners, imaging conditions, contrast injection protocols and local clinical practise. This work investigates for the first time multi-center and multi-vendor LV segmentation in LGE-MRI, by proposing, implementing and evaluating in detail several strategies to enhance model generalizability across clinical cites. These include data augmentation to artificially augment the image variability in the training sample, image harmonization to align the distributions of LGE-MRI images across centers, and transfer learning to adjust existing single-center models to unseen images from new clinical sites. The results obtained based on a new multi-center LGE-MRI dataset acquired in four clinical centers in Spain, France and China, show that the combination of data augmentation and transfer learning can lead to single-center models that generalize well to new clinical centers not included in the original training. The proposed framework shows the potential for developing clinical tools for automated LV segmentation in LGE-MRI that can be deployed in multiple clinical centers across distinct geographical locations.

(0 min) Video monitoring of traffic is useful for traffic management and control, traffic counting, and traffic law enforcement. However, traffic monitoring during inclement weather such as rain is a challenging task because video quality is corrupted by streaks of falling rain on the video image, and this hinders reliable characterization not only of the road environment but also of road-user behavior during such adverse weather events. This study proposes a two-stage self-supervised learning method to remove rain streaks in traffic videos. The first and second stages address intra- and inter-frame noise, respectively. The results indicated that the model exhibits satisfactory performance in terms of the image visual quality and the Peak Signal-Noise Ratio value.

(0 min) In this paper, we introduce a variational Bayesian algorithm (VBA) for image blind deconvolution. Our generic framework incorporates smoothness priors on the unknown blur/image and possible affine constraints (e.g., sum to one) on the blur kernel. One of our main contributions is the integration of VBA within a neural network paradigm, following an unrolling methodology. The proposed architecture is trained in a supervised fashion, which allows us to optimally set two key hyperparameters of the VBA model and lead to further improvements in terms of resulting visual quality. Various experiments involving grayscale/color images and diverse kernel shapes, are performed. The numerical examples illustrate the high performance of our approach when compared to state-of-the-art techniques based on optimization, Bayesian estimation, or deep learning.

(0 min) Image- and video-based 3D human recovery (i.e. pose and shape estimation) have achieved substantial progress. However, due to the prohibitive cost of motion capture, existing datasets are often limited in scale and diversity, which hinders the further development of more powerful models. In this work, we obtain massive human sequences as well as their 3D ground truths by playing video games. Specifically, we contribute, GTA-Human, a mega-scale and highly-diverse 3D human dataset generated with the GTA-V game engine. With a rich set of subjects, actions, and scenarios, GTA-Human serves as both an effective training source. Notably, the "unreasonable effectiveness of data" phenomenon is validated in 3D human recovery using our game-playing data. A simple frame-based baseline trained on GTA-Human already outperforms more sophisticated methods by a large margin; for video-based methods, GTA-Human demonstrates superiority over even the in-domain training set. We extend our study to larger models to observe the same consistent improvements, and the study on supervision signals suggests the rich collection of SMPL annotations is key. Furthermore, equipped with the diverse annotations in GTA-Human, we systematically investigate the performance of various methods under a wide spectrum of real-world variations, e.g. camera angles, poses, and occlusions. We hope our work could pave way for scaling up 3D human recovery to the real world.

(0 min) We present a relocalization pipeline, which combines an absolute pose regression (APR) network with a novel view synthesis based direct matching module, offering superior accuracy while maintaining low inference time. Our contribution is twofold: i) we design a direct matching module that supplies a photometric supervision signal to refine the pose regression network via differentiable rendering; ii) we modify the rotation representation from the classical quaternion to SO(3) in pose regression, removing the need for balancing rotation and translation loss terms. As a result, our network Direct-PoseNet achieves state-of-the-art performance among all other single-image APR methods on the 7-Scenes benchmark and the LLFF dataset.

(0 min) Global human motion forecasting is important in many fields, which is the combination of global human trajectory prediction and local human pose prediction. Visual and social information are often used to boost model performance, however, they may consume too much computational resource. In this paper, we establish a simple but effective baseline for single human motion forecasting without visual and social information, equipped with useful training tricks. Our method "futuremotion_ICCV21" outperforms existing methods by a large margin on SoMoF benchmark. We hope our work provide new ideas for future research.

(0 min) Multi-Task Learning (MTL) aims to enhance the model generalization by sharing representations between related tasks for better performance. Typical MTL methods are jointly trained with the complete multitude of ground-truths for all tasks simultaneously. However, one single dataset may not contain the annotations for each task of interest. To address this issue, we propose the Semi-supervised Multi-Task Learning (SemiMTL) method to leverage the available supervisory signals from different datasets, particularly for semantic segmentation and depth estimation tasks. To this end, we design an adversarial learning scheme in our semi-supervised training by leveraging unlabeled data to optimize all the task branches simultaneously and accomplish all tasks across datasets with partial annotations. We further present a domain-aware discriminator structure with various alignment formulations to mitigate the domain discrepancy issue among datasets. Finally, we demonstrate the effectiveness of the proposed method to learn across different datasets on challenging street view and remote sensing benchmarks.

(0 min) Central to the application of many multi-view geometry algorithms is the extraction of matching points between multiple viewpoints, enabling classical tasks such as camera pose estimation and 3D reconstruction. Many approaches that characterize these points have been proposed based on hand-tuned appearance models or data-driven learning methods. We propose Soft Expectation and Deep Maximization (SEDM), an iterative unsupervised learning process that directly optimizes the repeatability of the features by posing the problem in a similar way to expectation maximization (EM). We found convergence to be reliable and the new model to be more lighting invariant and better at localize the underlying 3D points in a scene, improving SfM quality when compared to other state of the art deep learning detectors.

(0 min) The current state-of-the-art methods for video corpus moment retrieval (VCMR) often use similarity-based feature alignment approach for the sake of convenience and speed. However, late fusion methods like cosine similarity alignment are unable to make full use of the information from both query texts and videos. In this paper, we combine feature alignment with feature fusion to promote the performance on VCMR.

(0 min) Contrastive self-supervised learning (CSL) has attracted increasing attention for model pre-training via unlabeled data. The resulted CSL models provide instance-discriminative visual features that are uniformly scattered in the feature space. During deployment, the common practice is to directly fine-tune CSL models with cross-entropy, which however may not be the best strategy in practice. Although cross-entropy tends to separate inter-class features, the resulting models still have limited capability for reducing intra-class feature scattering that exists in CSL models. In this paper, we investigate whether applying contrastive learning to fine-tuning would bring further benefits, and analytically find that optimizing the contrastive loss benefits both discriminative representation learning and model optimization during fine-tuning. Inspired by these findings, we propose Contrast-regularized tuning (Core-tuning), a new approach for fine-tuning CSL models. Instead of simply adding the contrastive loss to the objective of fine-tuning, Core-tuning further applies a novel hard pair mining strategy for more effective contrastive fine-tuning, as well as smoothing the decision boundary to better exploit the learned discriminative feature space. Extensive experiments on image classification and semantic segmentation verify the effectiveness of Core-tuning.

(0 min) We present TWIST, a novel self-supervised representation learning method by classifying large-scale unlabeled datasets in an end-to-end way. We employ a siamese network terminated by a softmax operation to produce twin class distributions of two augmented images. Without supervision, we enforce the class distributions of different augmentations to be consistent. In the meantime, we regularize the class distributions to make them sharp and diverse. Specifically, we minimize the entropy of the distribution for each sample to make the class prediction for each sample assertive and maximize the entropy of the mean distribution to make the predictions of different samples diverse. In this way, TWIST can naturally avoid the trivial solutions without specific designs such as asymmetric network, stop-gradient operation, or momentum encoder. Different from the clustering-based methods which alternate between clustering and learning, our method is a single learning process guided by a unified loss function. As a result, TWIST outperforms state-of-the-art methods on a wide range of tasks, including unsupervised classification, linear classification, semi-supervised learning, transfer learning, and some dense prediction tasks such as detection and segmentation.

(0 min) We present a simple, yet effective, approach for self-supervised 3D human pose estimation. Unlike the prior work, we explore the temporal information next to the multi-view self-supervision. During training, we rely on triangulating 2D body pose estimates of a multiple-view camera system. A temporal convolutional neural network is trained with the generated 3D ground-truth and the geometric multi-view consistency loss, imposing geometrical constraints on the predicted 3D body skeleton. During inference, our model receives a sequence of 2D body pose estimates from a single-view to predict the 3D body pose for each of them. An extensive evaluation shows that our method achieves state-of-the-art performance in the Human3.6M and MPI-INF-3DHP benchmarks. Our code and models are publicly available at \url{https://github.com/vru2020/TM_HPE/}.

(0 min) An unsupervised data-driven nuclei segmentation method for histology images, called CBM, is proposed in this work. CBM consists of three modules applied in a block-wise manner: 1) data-driven color transform for energy compaction and dimension reduction, 2) data-driven binarization, and 3) incorporation of geometric priors with morphological processing. CBM comes from the first letter of the three modules - "Color transform", "Binarization" and "Morphological processing". Experiments on the MoNuSeg dataset validate the effectiveness of the proposed CBM method. CBM outperforms all other unsupervised methods and offers a competitive standing among supervised models based on the Aggregated Jaccard Index (AJI) metric.

(0 min) We propose a novel method for creating adversarial examples. Instead of perturbing pixels, we use an encoder-decoder representation of the input image and perturb intermediate layers in the decoder. This changes the high-level features provided by the generative model. Therefore, our perturbation possesses semantic meaning, such as a longer beak or green tints. We formulate this task as an optimization problem by minimizing the Wasserstein distance between the adversarial and initial images under a misclassification constraint. We employ the projected gradient method with a simple inexact projection. Due to the projection, all iterations are feasible, and our method always generates adversarial images. We perform numerical experiments on the MNIST and ImageNet datasets in both targeted and untargeted settings. We demonstrate that our adversarial images are much less vulnerable to steganographic defence techniques than pixel-based attacks. Moreover, we show that our method modifies key features such as edges and that defence techniques based on adversarial training are vulnerable to our attacks.

(0 min) Equivariance has emerged as a desirable property of representations of objects subject to identity-preserving transformations that constitute a group, such as translations and rotations. However, the expressivity of a representation constrained by group equivariance is still not fully understood. We address this gap by providing a generalization of Cover's Function Counting Theorem that quantifies the number of linearly separable and group-invariant binary dichotomies that can be assigned to equivariant representations of objects. We find that the fraction of separable dichotomies is determined by the dimension of the space that is fixed by the group action. We show how this relation extends to operations such as convolutions, element-wise nonlinearities, and global and local pooling. While other operations do not change the fraction of separable dichotomies, local pooling decreases the fraction, despite being a highly nonlinear operation. Finally, we test our theory on intermediate representations of randomly initialized and fully trained convolutional neural networks and find perfect agreement.

(0 min) Long-term object detection requires the integration of frame-based results over several seconds. For non-deformable objects, long-term detection is often addressed using object detection followed by video tracking. Unfortunately, tracking is inapplicable to objects that undergo dramatic changes in appearance from frame to frame. As a related example, we study hand detection over long video recordings in collaborative learning environments. More specifically, we develop long-term hand detection methods that can deal with partial occlusions and dramatic changes in appearance. Our approach integrates object-detection, followed by time projections, clustering, and small region removal to provide effective hand detection over long videos. The hand detector achieved average precision (AP) of 72% at 0.5 intersection over union (IoU). The detection results were improved to 81% by using our optimized approach for data augmentation. The method runs at 4.7x the real-time with AP of 81% at 0.5 intersection over the union. Our method reduced the number of false-positive hand detections by 80% by improving IoU ratios from 0.2 to 0.5. The overall hand detection system runs at 4x real-time.

(0 min) This paper presents an efficient multi-scale vision Transformer, called ResT, that capably served as a general-purpose backbone for image recognition. Unlike existing Transformer methods, which employ standard Transformer blocks to tackle raw images with a fixed resolution, our ResT have several advantages: (1) A memory-efficient multi-head self-attention is built, which compresses the memory by a simple depth-wise convolution, and projects the interaction across the attention-heads dimension while keeping the diversity ability of multi-heads; (2) Position encoding is constructed as spatial attention, which is more flexible and can tackle with input images of arbitrary size without interpolation or fine-tune; (3) Instead of the straightforward tokenization at the beginning of each stage, we design the patch embedding as a stack of overlapping convolution operation with stride on the 2D-reshaped token map. We comprehensively validate ResT on image classification and downstream tasks. Experimental results show that the proposed ResT can outperform the recently state-of-the-art backbones by a large margin, demonstrating the potential of ResT as strong backbones. The code and models will be made publicly available at https://github.com/wofmanaf/ResT.

(0 min) Recent history has seen a tremendous growth of work exploring implicit representations of geometry and radiance, popularized through Neural Radiance Fields (NeRF). Such works are fundamentally based on a (implicit) {\em volumetric} representation of occupancy, allowing them to model diverse scene structure including translucent objects and atmospheric obscurants. But because the vast majority of real-world scenes are composed of well-defined surfaces, we introduce a {\em surface} analog of such implicit models called Neural Reflectance Surfaces (NeRS). NeRS learns a neural shape representation of a closed surface that is diffeomorphic to a sphere, guaranteeing water-tight reconstructions. Even more importantly, surface parameterizations allow NeRS to learn (neural) bidirectional surface reflectance functions (BRDFs) that factorize view-dependent appearance into environmental illumination, diffuse color (albedo), and specular "shininess." Finally, rather than illustrating our results on synthetic scenes or controlled in-the-lab capture, we assemble a novel dataset of multi-view images from online marketplaces for selling goods. Such "in-the-wild" multi-view image sets pose a number of challenges, including a small number of views with unknown/rough camera estimates. We demonstrate that surface-based neural reconstructions enable learning from such data, outperforming volumetric neural rendering-based reconstructions. We hope that NeRS serves as a first step toward building scalable, high-quality libraries of real-world shape, materials, and illumination. The project page with code and video visualizations can be found at https://jasonyzhang.com/ners}{jasonyzhang.com/ners.

(0 min) Weakly supervised object detection (WSOD) has attracted more and more attention since it only uses image-level labels and can save huge annotation costs. Most of the WSOD methods use Multiple Instance Learning (MIL) as their basic framework, which regard it as an instance classification problem. However, these methods based on MIL tends to converge only on the most discriminate regions of different instances, rather than their corresponding complete regions, that is, insufficient integrity. Inspired by the habit of observing things by the human, we propose a new method by comparing the initial proposals and the extension ones to optimize those initial proposals. Specifically, we propose one new strategy for WSOD by involving contrastive proposal extension (CPE), which consists of multiple directional contrastive proposal extensions (D-CPE), and each D-CPE contains encoders based on LSTM network and corresponding decoders. %\textcolor{red}{with temporal network}. Firstly, the boundary of initial proposals in MIL is extended to different positions according to well-designed sequential order. Then, CPE compares the extended proposal and the initial proposal by extracting the feature semantics of them using the encoders, and calculates the integrity of the initial proposal to optimize the score of the initial proposal.

(0 min) Visual localization is a crucial component in the application of mobile robot and autonomous driving. Image retrieval is an efficient and effective technique in image-based localization methods. Due to the drastic variability of environmental conditions, e.g. illumination, seasonal and weather changes, retrieval-based visual localization is severely affected and becomes a challenging problem. In this work, a general architecture is first formulated probabilistically to extract domain invariant feature through multi-domain image translation. And then a novel gradient-weighted similarity activation mapping loss (Grad-SAM) is incorporated for finer localization with high accuracy. We also propose a new adaptive triplet loss to boost the contrastive learning of the embedding in a self-supervised manner. The final coarse-to-fine image retrieval pipeline is implemented as the sequential combination of models without and with Grad-SAM loss. Extensive experiments have been conducted to validate the effectiveness of the proposed approach on the CMUSeasons dataset. The strong generalization ability of our approach is verified on RobotCar dataset using models pre-trained on urban part of CMU-Seasons dataset. Our performance is on par with or even outperforms the state-of-the-art image-based localization baselines in medium or high precision, especially under the challenging environments with illumination variance, vegetation and night-time images. The code and pretrained models are available on https://github.com/HanjiangHu/DISAM.

(0 min) In this work, we present Point Transformer, a deep neural network that operates directly on unordered and unstructured point sets. We design Point Transformer to extract local and global features and relate both representations by introducing the local-global attention mechanism, which aims to capture spatial point relations and shape information. For that purpose, we propose SortNet, as part of the Point Transformer, which induces input permutation invariance by selecting points based on a learned score. The output of Point Transformer is a sorted and permutation invariant feature list that can directly be incorporated into common computer vision applications. We evaluate our approach on standard classification and part segmentation benchmarks to demonstrate competitive results compared to the prior work. Code is publicly available at: https://github.com/engelnico/point-transformer

(0 min) Zero-shot learning (ZSL) aims to recognize unseen classes based on the knowledge of seen classes. Previous methods focused on learning direct embeddings from global features to the semantic space in hope of knowledge transfer from seen classes to unseen classes. However, an unseen class shares local visual features with a set of seen classes and leveraging global visual features makes the knowledge transfer ineffective. To tackle this problem, we propose a Region Semantically Aligned Network (RSAN), which maps local features of unseen classes to their semantic attributes. Instead of using global features which are obtained by an average pooling layer after an image encoder, we directly utilize the output of the image encoder which maintains local information of the image. Concretely, we obtain each attribute from a specific region of the output and exploit these attributes for recognition. As a result, the knowledge of seen classes can be successfully transferred to unseen classes in a region-bases manner. In addition, we regularize the image encoder through attribute regression with a semantic knowledge to extract robust and attribute-related visual features. Experiments on several standard ZSL datasets reveal the benefit of the proposed RSAN method, outperforming state-of-the-art methods.

(0 min) Learning to understand and predict future motions or behaviors for agents like humans and robots are critical to various autonomous platforms, such as behavior analysis, robot navigation, and self-driving cars. Intrinsic factors such as agents' diversified personalities and decision-making styles bring rich and diverse changes and multi-modal characteristics to their future plannings. Besides, the extrinsic interactive factors have also brought rich and varied changes to their trajectories. Previous methods mostly treat trajectories as time sequences, and reach great prediction performance. In this work, we try to focus on agents' trajectories in another view, i.e., the Fourier spectrums, to explore their future behavior rules in a novel hierarchical way. We propose the Transformer-based V model, which concatenates two continuous keypoints estimation and spectrum interpolation sub-networks, to model and predict agents' trajectories with spectrums in the keypoints and interactions levels respectively. Experimental results show that V outperforms most of current state-of-the-art methods on ETH-UCY and SDD trajectories dataset for about 15\% quantitative improvements, and performs better qualitative results.

(0 min) The goal of this paper is to learn strong lip reading models that can recognise speech in silent videos. Most prior works deal with the open-set visual speech recognition problem by adapting existing automatic speech recognition techniques on top of trivially pooled visual features. Instead, in this paper we focus on the unique challenges encountered in lip reading and propose tailored solutions. To that end we make the following contributions: (1) we propose an attention-based pooling mechanism to aggregate visual speech representations; (2) we use sub-word units for lip reading for the first time and show that this allows us to better model the ambiguities of the task; (3) we propose a training pipeline that balances the lip reading performance with other key factors such as data and compute efficiency. Following the above, we obtain state-of-the-art results on the challenging LRS2 and LRS3 benchmarks when training on public datasets, and even surpass models trained on large-scale industrial datasets by using an order of magnitude less data. Our best model achieves 22.6% word error rate on the LRS2 dataset, a performance unprecedented for lip reading models, significantly reducing the performance gap between lip reading and automatic speech recognition.

(0 min) Object detectors usually achieve promising results with the supervision of complete instance annotations. However, their performance is far from satisfactory with sparse instance annotations. Most existing methods for sparsely annotated object detection either re-weight the loss of hard negative samples or convert the unlabeled instances into ignored regions to reduce the interference of false negatives. We argue that these strategies are insufficient since they can at most alleviate the negative effect caused by missing annotations. In this paper, we propose a simple but effective mechanism, called Co-mining, for sparsely annotated object detection. In our Co-mining, two branches of a Siamese network predict the pseudo-label sets for each other. To enhance multi-view learning and better mine unlabeled instances, the original image and corresponding augmented image are used as the inputs of two branches of the Siamese network, respectively. Co-mining can serve as a general training mechanism applied to most of modern object detectors. Experiments are performed on MS COCO dataset with three different sparsely annotated settings using two typical frameworks: anchor-based detector RetinaNet and anchor-free detector FCOS. Experimental results show that our Co-mining with RetinaNet achieves 1.4%~2.1% improvements compared with different baselines and surpasses existing methods under the same sparsely annotated setting. Code is available at https://github.com/megvii-research/Co-mining.

(0 min) Drone-based vehicle detection aims at finding the vehicle locations and categories in an aerial image. It empowers smart city traffic management and disaster rescue. Researchers have made mount of efforts in this area and achieved considerable progress. Nevertheless, it is still a challenge when the objects are hard to distinguish, especially in low light conditions. To tackle this problem, we construct a large-scale drone-based RGB-Infrared vehicle detection dataset, termed DroneVehicle. Our DroneVehicle collects 28, 439 RGB-Infrared image pairs, covering urban roads, residential areas, parking lots, and other scenarios from day to night. Due to the great gap between RGB and infrared images, cross-modal images provide both effective information and redundant information. To address this dilemma, we further propose an uncertainty-aware cross-modality vehicle detection (UA-CMDet) framework to extract complementary information from cross-modal images, which can significantly improve the detection performance in low light conditions. An uncertainty-aware module (UAM) is designed to quantify the uncertainty weights of each modality, which is calculated by the cross-modal Intersection over Union (IoU) and the RGB illumination value. Furthermore, we design an illumination-aware cross-modal non-maximum suppression algorithm to better integrate the modal-specific information in the inference phase. Extensive experiments on the DroneVehicle dataset demonstrate the flexibility and effectiveness of the proposed method for crossmodality vehicle detection. The dataset can be download from https://github.com/VisDrone/DroneVehicle.

(0 min) We experimentally study the robustness of deep camera-LiDAR fusion architectures for 2D object detection in autonomous driving. First, we find that the fusion model is usually both more accurate, and more robust against single-source attacks than single-sensor deep neural networks. Furthermore, we show that without adversarial training, early fusion is more robust than late fusion, whereas the two perform similarly after adversarial training. However, we note that single-channel adversarial training of deep fusion is often detrimental even to robustness. Moreover, we observe cross-channel externalities, where single-channel adversarial training reduces robustness to attacks on the other channel. Additionally, we observe that the choice of adversarial model in adversarial training is critical: using attacks restricted to cars' bounding boxes is more effective in adversarial training and exhibits less significant cross-channel externalities. Finally, we find that joint-channel adversarial training helps mitigate many of the issues above, but does not significantly boost adversarial robustness.

(0 min) Typical learning-based light field reconstruction methods demand in constructing a large receptive field by deepening the network to capture correspondences between input views. In this paper, we propose a spatial-angular attention network to perceive correspondences in the light field non-locally, and reconstruction high angular resolution light field in an end-to-end manner. Motivated by the non-local attention mechanism, a spatial-angular attention module specifically for the high-dimensional light field data is introduced to compute the responses from all the positions in the epipolar plane for each pixel in the light field, and generate an attention map that captures correspondences along the angular dimension. We then propose a multi-scale reconstruction structure to efficiently implement the non-local attention in the low spatial scale, while also preserving the high frequency components in the high spatial scales. Extensive experiments demonstrate the superior performance of the proposed spatial-angular attention network for reconstructing sparsely-sampled light fields with non-Lambertian effects.

(0 min) We introduce a state-of-the-art audio-visual on-screen sound separation system which is capable of learning to separate sounds and associate them with on-screen objects by looking at in-the-wild videos. We identify limitations of previous work on audio-visual on-screen sound separation, including the simplicity and coarse resolution of spatio-temporal attention, and poor convergence of the audio separation model. Our proposed model addresses these issues using cross-modal and self-attention modules that capture audio-visual dependencies at a finer resolution over time, and by unsupervised pre-training of audio separation model. These improvements allow the model to generalize to a much wider set of unseen videos. We also show a robust way to further improve the generalization capability of our models by calibrating the probabilities of our audio-visual on-screen classifier, using only a small amount of in-domain videos labeled for their on-screen presence. For evaluation and semi-supervised training, we collected human annotations of on-screen audio from a large database of in-the-wild videos (YFCC100m). Our results show marked improvements in on-screen separation performance, in more general conditions than previous methods.

(0 min) Pre-training has enabled state-of-the-art results on many tasks. In spite of its recognized contribution to generalization, we observed in this study that pre-training also transfers adversarial non-robustness from pre-trained model into fine-tuned model in the downstream tasks. Using image classification as an example, we first conducted experiments on various datasets and network backbones to uncover the adversarial non-robustness in fine-tuned model. Further analysis was conducted on examining the learned knowledge of fine-tuned model and standard model, and revealed that the reason leading to the non-robustness is the non-robust features transferred from pre-trained model. Finally, we analyzed the preference for feature learning of the pre-trained model, explored the factors influencing robustness, and introduced a simple robust pre-traning solution.

(0 min) Analysis of vision-and-language models has revealed their brittleness under linguistic phenomena such as paraphrasing, negation, textual entailment, and word substitutions with synonyms or antonyms. While data augmentation techniques have been designed to mitigate against these failure modes, methods that can integrate this knowledge into the training pipeline remain under-explored. In this paper, we present \textbf{SDRO}, a model-agnostic method that utilizes a set linguistic transformations in a distributed robust optimization setting, along with an ensembling technique to leverage these transformations during inference. Experiments on benchmark datasets with images (NLVR$^2$) and video (VIOLIN) demonstrate performance improvements as well as robustness to adversarial attacks. Experiments on binary VQA explore the generalizability of this method to other V\&L tasks.

(0 min) Learning a good representation for space-time correspondence is the key for various computer vision tasks, including tracking object bounding boxes and performing video object pixel segmentation. To learn generalizable representation for correspondence in large-scale, a variety of self-supervised pretext tasks are proposed to explicitly perform object-level or patch-level similarity learning. Instead of following the previous literature, we propose to learn correspondence using Video Frame-level Similarity (VFS) learning, i.e, simply learning from comparing video frames. Our work is inspired by the recent success in image-level contrastive learning and similarity learning for visual recognition. Our hypothesis is that if the representation is good for recognition, it requires the convolutional features to find correspondence between similar objects or parts. Our experiments show surprising results that VFS surpasses state-of-the-art self-supervised approaches for both OTB visual object tracking and DAVIS video object segmentation. We perform detailed analysis on what matters in VFS and reveals new properties on image and frame level similarity learning. Project page with code is available at https://jerryxu.net/VFS

(0 min) Statistical shape modeling (SSM) characterizes anatomical variations in a population of shapes generated from medical images. SSM requires consistent shape representation across samples in shape cohort. Establishing this representation entails a processing pipeline that includes anatomy segmentation, re-sampling, registration, and non-linear optimization. These shape representations are then used to extract low-dimensional shape descriptors that facilitate subsequent analyses in different applications. However, the current process of obtaining these shape descriptors from imaging data relies on human and computational resources, requiring domain expertise for segmenting anatomies of interest. Moreover, this same taxing pipeline needs to be repeated to infer shape descriptors for new image data using a pre-trained/existing shape model. Here, we propose DeepSSM, a deep learning-based framework for learning the functional mapping from images to low-dimensional shape descriptors and their associated shape representations, thereby inferring statistical representation of anatomy directly from 3D images. Once trained using an existing shape model, DeepSSM circumvents the heavy and manual pre-processing and segmentation and significantly improves the computational time, making it a viable solution for fully end-to-end SSM applications. In addition, we introduce a model-based data-augmentation strategy to address data scarcity. Finally, this paper presents and analyzes two different architectural variants of DeepSSM with different loss functions using three medical datasets and their downstream clinical application. Experiments showcase that DeepSSM performs comparably or better to the state-of-the-art SSM both quantitatively and on application-driven downstream tasks. Therefore, DeepSSM aims to provide a comprehensive blueprint for deep learning-based image-to-shape models.

(0 min) Click-through rate prediction (CTR) and post-click conversion rate prediction (CVR) play key roles across all industrial ranking systems, such as recommendation systems, online advertising, and search engines. Different from the extensive research on CTR, there is much less research on CVR estimation, whose main challenge is extreme data sparsity with one or two orders of magnitude reduction in the number of samples than CTR. People try to solve this problem with the paradigm of multi-task learning with the sufficient samples of CTR, but the typical hard sharing method can't effectively solve this problem, because it is difficult to analyze which parts of network components can be shared and which parts are in conflict, i.e., there is a large inaccuracy with artificially designed neurons sharing. In this paper, we model CVR in a brand-new method by adopting the lottery-ticket-hypothesis-based sparse sharing multi-task learning, which can automatically and flexibly learn which neuron weights to be shared without artificial experience. Experiments on the dataset gathered from traffic logs of Tencent video's recommendation system demonstrate that sparse sharing in the CVR model significantly outperforms competitive methods. Due to the nature of weight sparsity in sparse sharing, it can also significantly reduce computational complexity and memory usage which are very important in the industrial recommendation system.

(0 min) It remains challenging to automatically predict the multi-agent trajectory due to multiple interactions including agent to agent interaction and scene to agent interaction. Although recent methods have achieved promising performance, most of them just consider spatial influence of the interactions and ignore the fact that temporal influence always accompanies spatial influence. Moreover, those methods based on scene information always require extra segmented scene images to generate multiple socially acceptable trajectories. To solve these limitations, we propose a novel model named spatial-temporal attentive network with spatial continuity (STAN-SC). First, spatial-temporal attention mechanism is presented to explore the most useful and important information. Second, we conduct a joint feature sequence based on the sequence and instant state information to make the generative trajectories keep spatial continuity. Experiments are performed on the two widely used ETH-UCY datasets and demonstrate that the proposed model achieves state-of-the-art prediction accuracy and handles more complex scenarios.

(0 min) Existing position based point cloud filtering methods can hardly preserve sharp geometric features. In this paper, we rethink point cloud filtering from a non-learning non-local non-normal perspective, and propose a novel position based approach for feature-preserving point cloud filtering. Unlike normal based techniques, our method does not require the normal information. The core idea is to first design a similarity metric to search the non-local similar patches of a queried local patch. We then map the non-local similar patches into a canonical space and aggregate the non-local information. The aggregated outcome (i.e. coordinate) will be inversely mapped into the original space. Our method is simple yet effective. Extensive experiments validate our method, and show that it generally outperforms position based methods (deep learning and non-learning), and generates better or comparable outcomes to normal based techniques (deep learning and non-learning).

(0 min) In recent years, unsupervised domain adaptation (UDA) for semantic segmentation has brought many researchers'attention. Many of them take an approach to design a complex system so as to better align the gap between source and target domain. Instead, we focus on the very basic structure of the deep neural network, Batch Normalization, and propose to replace the Sharing Affine Transformation with our proposed Separate Affine Transformation (SEAT). The proposed SEAT is simple, easily implemented and easy to integrate into existing adversarial learning based UDA methods. Also, to further improve the adaptation quality, we introduce multi level adaptation by adding the lower-level features to the higher-level ones before feeding them to the discriminator, without adding extra discriminator like others. Experiments show that the proposed methods is less complex without losing performance accuracy when compared with other UDA methods.

(0 min) Visually-grounded spoken language datasets can enable models to learn cross-modal correspondences with very weak supervision. However, modern audio-visual datasets contain biases that undermine the real-world performance of models trained on that data. We introduce Spoken ObjectNet, which is designed to remove some of these biases and provide a way to better evaluate how effectively models will perform in real-world scenarios. This dataset expands upon ObjectNet, which is a bias-controlled image dataset that features similar image classes to those present in ImageNet. We detail our data collection pipeline, which features several methods to improve caption quality, including automated language model checks. Lastly, we show baseline results on image retrieval and audio retrieval tasks. These results show that models trained on other datasets and then evaluated on Spoken ObjectNet tend to perform poorly due to biases in other datasets that the models have learned. We also show evidence that the performance decrease is due to the dataset controls, and not the transfer setting.

(0 min) Modern works on style transfer focus on transferring style from a single image. Recently, some approaches study multiple style transfer; these, however, are either too slow or fail to mix multiple styles. We propose ST-VAE, a Variational AutoEncoder for latent space-based style transfer. It performs multiple style transfer by projecting nonlinear styles to a linear latent space, enabling to merge styles via linear interpolation before transferring the new style to the content image. To evaluate ST-VAE, we experiment on COCO for single and multiple style transfer. We also present a case study revealing that ST-VAE outperforms other methods while being faster, flexible, and setting a new path for multiple style transfer.

(0 min) Diminished reality is a technology that aims to remove objects from video images and fills in the missing region with plausible pixels. Most conventional methods utilize the different cameras that capture the same scene from different viewpoints to allow regions to be removed and restored. In this paper, we propose an RGB-D image inpainting method using generative adversarial network, which does not require multiple cameras. Recently, an RGB image inpainting method has achieved outstanding results by employing a generative adversarial network. However, RGB inpainting methods aim to restore only the texture of the missing region and, therefore, does not recover geometric information (i.e, 3D structure of the scene). We expand conventional image inpainting method to RGB-D image inpainting to jointly restore the texture and geometry of missing regions from a pair of RGB and depth images. Inspired by other tasks that use RGB and depth images (e.g., semantic segmentation and object detection), we propose late fusion approach that exploits the advantage of RGB and depth information each other. The experimental results verify the effectiveness of our proposed method.

(0 min) Most approaches to camera calibration rely on calibration targets of well-known geometry. During data acquisition, calibration target and camera system are typically moved w.r.t. each other, to allow image coverage and perspective versatility. We show that moving the target can lead to small temporary deformations of the target, which can introduce significant errors into the calibration result. While static inaccuracies of calibration targets have been addressed in previous works, to our knowledge, none of the existing approaches can capture time-varying, dynamic deformations. To achieve high-accuracy calibrations despite moving the target, we propose a way to explicitly model dynamic target deformations in camera calibration. This is achieved by using a low-dimensional deformation model with only few parameters per image, which can be optimized jointly with target poses and intrinsics. We demonstrate the effectiveness of modeling dynamic deformations using different calibration targets and show its significance in a structure-from-motion application.

(0 min) We introduce HUMAN4D, a large and multimodal 4D dataset that contains a variety of human activities simultaneously captured by a professional marker-based MoCap, a volumetric capture and an audio recording system. By capturing 2 female and $2$ male professional actors performing various full-body movements and expressions, HUMAN4D provides a diverse set of motions and poses encountered as part of single- and multi-person daily, physical and social activities (jumping, dancing, etc.), along with multi-RGBD (mRGBD), volumetric and audio data. Despite the existence of multi-view color datasets captured with the use of hardware (HW) synchronization, to the best of our knowledge, HUMAN4D is the first and only public resource that provides volumetric depth maps with high synchronization precision due to the use of intra- and inter-sensor HW-SYNC. Moreover, a spatio-temporally aligned scanned and rigged 3D character complements HUMAN4D to enable joint research on time-varying and high-quality dynamic meshes. We provide evaluation baselines by benchmarking HUMAN4D with state-of-the-art human pose estimation and 3D compression methods. For the former, we apply 2D and 3D pose estimation algorithms both on single- and multi-view data cues. For the latter, we benchmark open-source 3D codecs on volumetric data respecting online volumetric video encoding and steady bit-rates. Furthermore, qualitative and quantitative visual comparison between mesh-based volumetric data reconstructed in different qualities showcases the available options with respect to 4D representations. HUMAN4D is introduced to the computer vision and graphics research communities to enable joint research on spatio-temporally aligned pose, volumetric, mRGBD and audio data cues. The dataset and its code are available https://tofis.github.io/myurls/human4d.

(0 min) Social concepts referring to non-physical objects--such as revolution, violence, or friendship--are powerful tools to describe, index, and query the content of visual data, including ever-growing collections of art images from the Cultural Heritage (CH) field. While much progress has been made towards complete image understanding in computer vision, automatic detection of social concepts evoked by images is still a challenge. This is partly due to the well-known semantic gap problem, worsened for social concepts given their lack of unique physical features, and reliance on more unspecific features than concrete concepts. In this paper, we propose the translation of recent cognitive theories about social concept representation into a software approach to represent them as multimodal frames, by integrating multisensory data. Our method focuses on the extraction, analysis, and integration of multimodal features from visual art material tagged with the concepts of interest. We define a conceptual model and present a novel ontology for formally representing social concepts as multimodal frames. Taking the Tate Gallery's collection as an empirical basis, we experiment our method on a corpus of art images to provide a proof of concept of its potential. We discuss further directions of research, and provide all software, data sources, and results.

(2 min) This study aims to explore different pre-trained models offered in the Torchvision package which is available in the PyTorch library. And investigate their effectiveness on fine-grained images classification. Transfer Learning is an effective method of achieving extremely good performance with insufficient training data. In many real-world situations, people cannot collect sufficient data required to train a deep neural network model efficiently. Transfer Learning models are pre-trained on a large data set, and can bring a good performance on smaller datasets with significantly lower training time. Torchvision package offers us many models to apply the Transfer Learning on smaller datasets. Therefore, researchers may need a guideline for the selection of a good model. We investigate Torchvision pre-trained models on four different data sets: 10 Monkey Species, 225 Bird Species, Fruits 360, and Oxford 102 Flowers. These data sets have images of different resolutions, class numbers, and different achievable accuracies. We also apply their usual fully-connected layer and the Spinal fully-connected layer to investigate the effectiveness of SpinalNet. The Spinal fully-connected layer brings better performance in most situations. We apply the same augmentation for different models for the same data set for a fair comparison. This paper may help future Computer Vision researchers in choosing a proper Transfer Learning model.

(2 min) In this paper, we present a Nuisance-label Supervision (NLS) module, which can make models more robust to nuisance factor variations. Nuisance factors are those irrelevant to a task, and an ideal model should be invariant to them. For example, an activity recognition model should perform consistently regardless of the change of clothes and background. But our experiments show existing models are far from this capability. So we explicitly supervise a model with nuisance labels to make extracted features less dependent on nuisance factors. Although the values of nuisance factors are rarely annotated, we demonstrate that besides existing annotations, nuisance labels can be acquired freely from data augmentation and synthetic data. Experiments show consistent improvement in robustness towards image corruption and appearance change in action recognition.

(2 min) Methods for model explainability have become increasingly critical for testing the fairness and soundness of deep learning. A number of explainability techniques have been developed which use a set of examples to represent a human-interpretable concept in a model's activations. In this work we show that these explainability methods can suffer the same vulnerability to adversarial attacks as the models they are meant to analyze. We demonstrate this phenomenon on two well-known concept-based approaches to the explainability of deep learning models: TCAV and faceted feature visualization. We show that by carefully perturbing the examples of the concept that is being investigated, we can radically change the output of the interpretability method, e.g. showing that stripes are not an important factor in identifying images of a zebra. Our work highlights the fact that in safety-critical applications, there is need for security around not only the machine learning pipeline but also the model interpretation process.

(2 min) Though image-level weakly supervised semantic segmentation (WSSS) has achieved great progress with Class Activation Map (CAM) as the cornerstone, the large supervision gap between classification and segmentation still hampers the model to generate more complete and precise pseudo masks for segmentation. In this study, we explore two implicit but intuitive constraints, i.e., cross-view feature semantic consistency and intra(inter)-class compactness(dispersion), to narrow the supervision gap. To this end, we propose two novel pixel-to-prototype contrast regularization terms that are conducted cross different views and within per single view of an image, respectively. Besides, we adopt two sample mining strategies, named semi-hard prototype mining and hard pixel sampling, to better leverage hard examples while reducing incorrect contrasts caused due to the absence of precise pixel-wise labels. Our method can be seamlessly incorporated into existing WSSS models without any changes to the base network and does not incur any extra inference burden. Experiments on standard benchmark show that our method consistently improves two strong baselines by large margins, demonstrating the effectiveness of our method. Specifically, built on top of SEAM, we improve the initial seed mIoU on PASCAL VOC 2012 from 55.4% to 61.5%. Moreover, armed with our method, we increase the segmentation mIoU of EPS from 70.8% to 73.6%, achieving new state-of-the-art.

(2 min) We demonstrate a working prototype for the monitoring of cow welfare by automatically analysing the animal behaviours. Deep learning models have been developed and tested with videos acquired in a farm, and a precision of 81.2\% has been achieved for cow identification. An accuracy of 84.4\% has been achieved for the detection of drinking events, and 94.4\% for the detection of grazing events. Experimental results show that the proposed deep learning method can be used to identify the behaviours of individual animals to enable automated farm provenance. Our raw and ground-truth dataset will be released as the first public video dataset for cow identification and action recognition. Recommendations for further development are also provided.

(2 min) We present SGoLAM, short for simultaneous goal localization and mapping, which is a simple and efficient algorithm for Multi-Object Goal navigation. Given an agent equipped with an RGB-D camera and a GPS/Compass sensor, our objective is to have the agent navigate to a sequence of target objects in realistic 3D environments. Our pipeline fully leverages the strength of classical approaches for visual navigation, by decomposing the problem into two key components: mapping and goal localization. The mapping module converts the depth observations into an occupancy map, and the goal localization module marks the locations of goal objects. The agent's policy is determined using the information provided by the two modules: if a current goal is found, plan towards the goal and otherwise, perform exploration. As our approach does not require any training of neural networks, it could be used in an off-the-shelf manner, and amenable for fast generalization in new, unseen environments. Nonetheless, our approach performs on par with the state-of-the-art learning-based approaches. SGoLAM is ranked 2nd in the CVPR 2021 MultiON (Multi-Object Goal Navigation) challenge. We have made our code publicly available at \emph{https://github.com/eunsunlee/SGoLAM}.

(2 min) Continual learning (CL) aims to learn a sequence of tasks without forgetting the previously acquired knowledge. However, recent advances in continual learning are restricted to supervised continual learning (SCL) scenarios. Consequently, they are not scalable to real-world applications where the data distribution is often biased and unannotated. In this work, we focus on unsupervised continual learning (UCL), where we learn the feature representations on an unlabelled sequence of tasks and show that reliance on annotated data is not necessary for continual learning. We conduct a systematic study analyzing the learned feature representations and show that unsupervised visual representations are surprisingly more robust to catastrophic forgetting, consistently achieve better performance, and generalize better to out-of-distribution tasks than SCL. Furthermore, we find that UCL achieves a smoother loss landscape through qualitative analysis of the learned representations and learns meaningful feature representations. Additionally, we propose Lifelong Unsupervised Mixup (LUMP), a simple yet effective technique that leverages the interpolation between the current task and previous tasks' instances to alleviate catastrophic forgetting for unsupervised representations.

(2 min) Contrastive learning of auditory and visual perception has been extremely successful when investigated individually. However, there are still major questions on how we could integrate principles learned from both domains to attain effective audiovisual representations. In this paper, we present a contrastive framework to learn audiovisual representations from unlabeled videos. The type and strength of augmentations utilized during self-supervised pre-training play a crucial role for contrastive frameworks to work sufficiently. Hence, we extensively investigate composition of temporal augmentations suitable for learning audiovisual representations; we find lossy spatio-temporal transformations that do not corrupt the temporal coherency of videos are the most effective. Furthermore, we show that the effectiveness of these transformations scales with higher temporal resolution and stronger transformation intensity. Compared to self-supervised models pre-trained on only sampling-based temporal augmentation, self-supervised models pre-trained with our temporal augmentations lead to approximately 6.5% gain on linear classifier performance on AVE dataset. Lastly, we show that despite their simplicity, our proposed transformations work well across self-supervised learning frameworks (SimSiam, MoCoV3, etc), and benchmark audiovisual dataset (AVE).

(3 min) We introduce Ego4D, a massive-scale egocentric video dataset and benchmark suite. It offers 3,025 hours of daily-life activity video spanning hundreds of scenarios (household, outdoor, workplace, leisure, etc.) captured by 855 unique camera wearers from 74 worldwide locations and 9 different countries. The approach to collection is designed to uphold rigorous privacy and ethics standards with consenting participants and robust de-identification procedures where relevant. Ego4D dramatically expands the volume of diverse egocentric video footage publicly available to the research community. Portions of the video are accompanied by audio, 3D meshes of the environment, eye gaze, stereo, and/or synchronized videos from multiple egocentric cameras at the same event. Furthermore, we present a host of new benchmark challenges centered around understanding the first-person visual experience in the past (querying an episodic memory), present (analyzing hand-object manipulation, audio-visual conversation, and social interactions), and future (forecasting activities). By publicly sharing this massive annotated dataset and benchmark suite, we aim to push the frontier of first-person perception. Project page: https://ego4d-data.org/

(2 min) The beet cyst nematode (BCN) Heterodera schachtii is a plant pest responsible for crop loss on a global scale. Here, we introduce a high-throughput system based on computer vision that allows quantifying BCN infestation and characterizing nematode cysts through phenotyping. After recording microscopic images of soil extracts in a standardized setting, an instance segmentation algorithm serves to detect nematode cysts in these samples. Going beyond fast and precise cyst counting, the image-based approach enables quantification of cyst density and phenotyping of morphological features of cysts under different conditions, providing the basis for high-throughput applications in agriculture and plant breeding research.

(2 min) In this paper, we propose a new deep unfolding neural network based on the ADMM algorithm for analysis Compressed Sensing. The proposed network jointly learns a redundant analysis operator for sparsification and reconstructs the signal of interest. We compare our proposed network with a state-of-the-art unfolded ISTA decoder, that also learns an orthogonal sparsifier. Moreover, we consider not only image, but also speech datasets as test examples. Computational experiments demonstrate that our proposed network outperforms the state-of-the-art deep unfolding networks, consistently for both real-world image and speech datasets.

(2 min) Few-shot class incremental learning -- the problem of updating a trained classifier to discriminate among an expanded set of classes with limited labeled data -- is a key challenge for machine learning systems deployed in non-stationary environments. Existing approaches to the problem rely on complex model architectures and training procedures that are difficult to tune and re-use. In this paper, we present an extremely simple approach that enables the use of ordinary logistic regression classifiers for few-shot incremental learning. The key to this approach is a new family of subspace regularization schemes that encourage weight vectors for new classes to lie close to the subspace spanned by the weights of existing classes. When combined with pretrained convolutional feature extractors, logistic regression models trained with subspace regularization outperform specialized, state-of-the-art approaches to few-shot incremental image classification by up to 22% on the miniImageNet dataset. Because of its simplicity, subspace regularization can be straightforwardly extended to incorporate additional background information about the new classes (including class names and descriptions specified in natural language); these further improve accuracy by up to 2%. Our results show that simple geometric regularization of class representations offers an effective tool for continual learning.

(2 min) Kinship verification is the task of determining whether a parent-child, sibling, or grandparent-grandchild relationship exists between two people and is important in social media applications, forensic investigations, finding missing children, and reuniting families. We demonstrate high quality kinship verification by participating in the FG 2021 Recognizing Families in the Wild challenge which provides the largest publicly available dataset in the field. Our approach is among the top 3 winning entries in the competition. We ensemble models written by both human experts and OpenAI Codex. We make our models and code publicly available.

(2 min) In this paper, we demonstrate how a generative model can be used to build a better recognizer through the control of content and style. We are building an online handwriting recognizer from a modest amount of training samples. By training our controllable handwriting synthesizer on the same data, we can synthesize handwriting with previously underrepresented content (e.g., URLs and email addresses) and style (e.g., cursive and slanted). Moreover, we propose a framework to analyze a recognizer that is trained with a mixture of real and synthetic training data. We use the framework to optimize data synthesis and demonstrate significant improvement on handwriting recognition over a model trained on real data only. Overall, we achieve a 66% reduction in Character Error Rate.

(2 min) How to robustly rank the aesthetic quality of given images has been a long-standing ill-posed topic. Such challenge stems mainly from the diverse subjective opinions of different observers about the varied types of content. There is a growing interest in estimating the user agreement by considering the standard deviation of the scores, instead of only predicting the mean aesthetic opinion score. Nevertheless, when comparing a pair of contents, few studies consider how confident are we regarding the difference in the aesthetic scores. In this paper, we thus propose (1) a re-adapted multi-task attention network to predict both the mean opinion score and the standard deviation in an end-to-end manner; (2) a brand-new confidence interval ranking loss that encourages the model to focus on image-pairs that are less certain about the difference of their aesthetic scores. With such loss, the model is encouraged to learn the uncertainty of the content that is relevant to the diversity of observers' opinions, i.e., user disagreement. Extensive experiments have demonstrated that the proposed multi-task aesthetic model achieves state-of-the-art performance on two different types of aesthetic datasets, i.e., AVA and TMGA.

(2 min) This technical report summarizes our method for the Video-And-Language Understanding Evaluation (VALUE) challenge (https://value-benchmark.github.io/challenge\_2021.html). We propose a CLIP-Enhanced method to incorporate the image-text pretrained knowledge into downstream video-text tasks. Combined with several other improved designs, our method outperforms the state-of-the-art by $2.4\%$ ($57.58$ to $60.00$) Meta-Ave score on VALUE benchmark.

(2 min) Dense retrieval (DR) methods conduct text retrieval by first encoding texts in the embedding space and then matching them by nearest neighbor search. This requires strong locality properties from the representation space, i.e, the close allocations of each small group of relevant texts, which are hard to generalize to domains without sufficient training data. In this paper, we aim to improve the generalization ability of DR models from source training domains with rich supervision signals to target domains without any relevant labels, in the zero-shot setting. To achieve that, we propose Momentum adversarial Domain Invariant Representation learning (MoDIR), which introduces a momentum method in the DR training process to train a domain classifier distinguishing source versus target, and then adversarially updates the DR encoder to learn domain invariant representations. Our experiments show that MoDIR robustly outperforms its baselines on 10+ ranking datasets from the BEIR benchmark in the zero-shot setup, with more than 10% relative gains on datasets with enough sensitivity for DR models' evaluation. Source code of this paper will be released.

(2 min) One of the challenges in information retrieval (IR) is the vocabulary mismatch problem, which happens when the terms between queries and documents are lexically different but semantically similar. While recent work has proposed to expand the queries or documents by enriching their representations with additional relevant terms to address this challenge, they usually require a large volume of query-document pairs to train an expansion model. In this paper, we propose an Unsupervised Document Expansion with Generation (UDEG) framework with a pre-trained language model, which generates diverse supplementary sentences for the original document without using labels on query-document pairs for training. For generating sentences, we further stochastically perturb their embeddings to generate more diverse sentences for document expansion. We validate our framework on two standard IR benchmark datasets. The results show that our framework significantly outperforms relevant expansion baselines for IR.

(2 min) Click-through rate prediction (CTR) and post-click conversion rate prediction (CVR) play key roles across all industrial ranking systems, such as recommendation systems, online advertising, and search engines. Different from the extensive research on CTR, there is much less research on CVR estimation, whose main challenge is extreme data sparsity with one or two orders of magnitude reduction in the number of samples than CTR. People try to solve this problem with the paradigm of multi-task learning with the sufficient samples of CTR, but the typical hard sharing method can't effectively solve this problem, because it is difficult to analyze which parts of network components can be shared and which parts are in conflict, i.e., there is a large inaccuracy with artificially designed neurons sharing. In this paper, we model CVR in a brand-new method by adopting the lottery-ticket-hypothesis-based sparse sharing multi-task learning, which can automatically and flexibly learn which neuron weights to be shared without artificial experience. Experiments on the dataset gathered from traffic logs of Tencent video's recommendation system demonstrate that sparse sharing in the CVR model significantly outperforms competitive methods. Due to the nature of weight sparsity in sparse sharing, it can also significantly reduce computational complexity and memory usage which are very important in the industrial recommendation system.

(2 min) Due to the development of graph neural networks, graph-based representation learning methods have made great progress in recommender systems. However, data sparsity is still a challenging problem that most graph-based recommendation methods are confronted with. Recent works try to address this problem by utilizing side information. In this paper, we exploit the relevant and easily accessible textual information by advanced natural language processing (NLP) models and propose a light RGCN-based (RGCN, relational graph convolutional network) collaborative filtering method based on heterogeneous graphs. Specifically, to incorporate rich textual knowledge, we utilize a pre-trained NLP model to initialize the embeddings of text nodes. Afterward, by performing a simplified RGCN-based node information propagation on the constructed heterogeneous graph, the embeddings of users and items can be adjusted with textual knowledge, which effectively alleviates the negative effects of data sparsity. Moreover, the matching function used by most graph-based representation learning methods is the inner product, which is not appropriate for the obtained embeddings that contain complex semantics. We design a predictive network that combines graph-based representation learning with neural matching function learning, and demonstrate that this architecture can bring a significant performance improvement. Extensive experiments are conducted on three publicly available datasets, and the results verify the superior performance of our method over several baselines.

(2 min) Food-choices and eating-habits directly contribute to our long-term health. This makes the food recommender system a potential tool to address the global crisis of obesity and malnutrition. Over the past decade, artificial-intelligence and medical researchers became more invested in researching tools that can guide and help people make healthy and thoughtful decisions around food and diet. In many typical (Recommender System) RS domains, smart nudges have been proven effective in shaping users' consumption patterns. In recent years, knowledgeable nudging and incentifying choices started getting attention in the food domain as well. To develop smart nudging for promoting healthier food choices, we combined Machine Learning and RS technology with food-healthiness guidelines from recognized health organizations, such as the World Health Organization, Food Standards Agency, and the National Health Service United Kingdom. In this paper, we discuss our research on, persuasive visualization for making users aware of the healthiness of the recommended recipes. Here, we propose three novel nudging technology, the WHO-BubbleSlider, the FSA-ColorCoading, and the DRCI-MLCP, that encourage users to choose healthier recipes. We also propose a Topic Modeling based portion-size recommendation algorithm. To evaluate our proposed smart-nudges, we conducted an online user study with 96 participants and 92250 recipes. Results showed that, during the food decision-making process, appropriate healthiness cues make users more likely to click, browse, and choose healthier recipes over less healthy ones.

(2 min) User profiling has long been an important problem that investigates user interests in many real applications. Some recent works regard users and their interacted objects as entities of a graph and turn the problem into a node classification task. However, they neglect the difference of distinct interaction types, e.g. user clicks an item v.s.user purchases an item, and thus cannot incorporate such information well. To solve these issues, we propose to leverage the relation-aware heterogeneous graph method for user profiling, which also allows capturing significant meta relations. We adopt the query, key, and value mechanism in a transformer fashion for heterogeneous message passing so that entities can effectively interact with each other. Via such interactions on different relation types, our model can generate representations with rich information for the user profile prediction. We conduct experiments on two real-world e-commerce datasets and observe a significant performance boost of our approach.

(2 min) While search technologies have evolved to be robust and ubiquitous, the fundamental interaction paradigm has remained relatively stable for decades. With the maturity of the Brain-Machine Interface, we build an efficient and effective communication system between human beings and search engines based on electroencephalogram~(EEG) signals, called Brain-Machine Search Interface(BMSI) system. The BMSI system provides functions including query reformulation and search result interaction. In our system, users can perform search tasks without having to use the mouse and keyboard. Therefore, it is useful for application scenarios in which hand-based interactions are infeasible, e.g, for users with severe neuromuscular disorders. Besides, based on brain signals decoding, our system can provide abundant and valuable user-side context information(e.g., real-time satisfaction feedback, extensive context information, and a clearer description of information needs) to the search engine, which is hard to capture in the previous paradigm. In our implementation, the system can decode user satisfaction from brain signals in real-time during the interaction process and re-rank the search results list based on user satisfaction feedback. The demo video is available at this http URL

(2 min) Recommender systems may be confounded by various types of confounding factors (also called confounders) that may lead to inaccurate recommendations and sacrificed recommendation performance. Current approaches to solving the problem usually design each specific model for each specific confounder. However, real-world systems may include a huge number of confounders and thus designing each specific model for each specific confounder is unrealistic. More importantly, except for those "explicit confounders" that researchers can manually identify and process such as item's position in the ranking list, there are also many "latent confounders" that are beyond the imagination of researchers. For example, users' rating on a song may depend on their current mood or the current weather, and users' preference on ice creams may depend on the air temperature. Such latent confounders may be unobservable in the recorded training data. To solve the problem, we propose a deconfounded causal collaborative filtering model. We first frame user behaviors with unobserved confounders into a causal graph, and then we design a front-door adjustment model carefully fused with machine learning to deconfound the influence of unobserved confounders. The proposed model is able to handle both global confounders and personalized confounders. Experiments on real-world e-commerce datasets show that our method is able to deconfound unobserved confounders to achieve better recommendation performance.

(2 min) With the growth of the academic engines, the mining and analysis acquisition of massive researcher data, such as collaborator recommendation and researcher retrieval, has become indispensable. It can improve the quality of services and intelligence of academic engines. Most of the existing studies for researcher data mining focus on a single task for a particular application scenario and learning a task-specific model, which is usually unable to transfer to out-of-scope tasks. The pre-training technology provides a generalized and sharing model to capture valuable information from enormous unlabeled data. The model can accomplish multiple downstream tasks via a few fine-tuning steps. In this paper, we propose a multi-task self-supervised learning-based researcher data pre-training model named RPT. Specifically, we divide the researchers' data into semantic document sets and community graph. We design the hierarchical Transformer and the local community encoder to capture information from the two categories of data, respectively. Then, we propose three self-supervised learning objectives to train the whole model. Finally, we also propose two transfer modes of RPT for fine-tuning in different scenarios. We conduct extensive experiments to evaluate RPT, results on three downstream tasks verify the effectiveness of pre-training for researcher data mining.

(2 min) Since late 2019, COVID-19 has quickly emerged as the newest biomedical domain, resulting in a surge of new information. As with other emergent domains, the discussion surrounding the topic has been rapidly changing, leading to the spread of misinformation. This has created the need for a public space for users to ask questions and receive credible, scientific answers. To fulfill this need, we turn to the task of open-domain question-answering, which we can use to efficiently find answers to free-text questions from a large set of documents. In this work, we present such a system for the emergent domain of COVID-19. Despite the small data size available, we are able to successfully train the system to retrieve answers from a large-scale corpus of published COVID-19 scientific papers. Furthermore, we incorporate effective re-ranking and question-answering techniques, such as document diversity and multiple answer spans. Our open-domain question-answering system can further act as a model for the quick development of similar systems that can be adapted and modified for other developing emergent domains.

(2 min) In this paper, we propose a new deep unfolding neural network based on the ADMM algorithm for analysis Compressed Sensing. The proposed network jointly learns a redundant analysis operator for sparsification and reconstructs the signal of interest. We compare our proposed network with a state-of-the-art unfolded ISTA decoder, that also learns an orthogonal sparsifier. Moreover, we consider not only image, but also speech datasets as test examples. Computational experiments demonstrate that our proposed network outperforms the state-of-the-art deep unfolding networks, consistently for both real-world image and speech datasets.

(2 min) Many legal professionals think that the explosion of information about local, regional, national, and international legislation makes their practice more costly, time-consuming, and even error-prone. The two main reasons for this are that most legislation is usually unstructured, and the tremendous amount and pace with which laws are released causes information overload in their daily tasks. In the case of the legal domain, the research community agrees that a system allowing to generate automatic responses to legal questions could substantially impact many practical implications in daily activities. The degree of usefulness is such that even a semi-automatic solution could significantly help to reduce the workload to be faced. This is mainly because a Question Answering system could be able to automatically process a massive amount of legal resources to answer a question or doubt in seconds, which means that it could save resources in the form of effort, money, and time to many professionals in the legal sector. In this work, we quantitatively and qualitatively survey the solutions that currently exist to meet this challenge.

(3 min) Movie-making has become one of the most costly and risky endeavors in the entertainment industry. Continuous change in the preference of the audience makes it harder to predict what kind of movie will be financially successful at the box office. So, it is no wonder that cautious, intelligent stakeholders and large production houses will always want to know the probable revenue that will be generated by a movie before making an investment. Researchers have been working on finding an optimal strategy to help investors in making the right decisions. But the lack of a large, up-to-date dataset makes their work harder. In this work, we introduce an up-to-date, richer, and larger dataset that we have prepared by scraping IMDb for researchers and data analysts to work with. The compiled dataset contains the summery data of 7.5 million titles and detail information of more than 200K movies. Additionally, we perform different statistical analysis approaches on our dataset to find out how a movie's revenue is affected by different pre-released attributes such as budget, runtime, release month, content rating, genre etc. In our analysis, we have found that having a star cast/director has a positive impact on generated revenue. We introduce a novel approach for calculating the star power of a movie. Based on our analysis we select a set of attributes as features and train different machine learning algorithms to predict a movie's expected revenue. Based on generated revenue, we classified the movies in 10 categories and achieved a one-class-away accuracy rate of almost 60% (bingo accuracy of 30%). All the generated datasets and analysis codes are available online. We also made the source codes of our scraper bots public, so that researchers interested in extending this work can easily modify these bots as they need and prepare their own up-to-date datasets.

(2 min) The essential task of Topic Detection and Tracking (TDT) is to organize a collection of news media into clusters of stories that pertain to the same real-world event. To apply TDT models to practical applications such as search engines and discovery tools, human guidance is needed to pin down the scope of an "event" for the corpus of interest. In this work in progress, we explore a human-in-the-loop method that helps users iteratively fine-tune TDT algorithms so that both the algorithms and the users themselves better understand the nature of the events. We generate a visual overview of the entire corpus, allowing the user to select regions of interest from the overview, and then ask a series of questions to affirm (or reject) that the selected documents belong to the same event. The answers to these questions supplement the training data for the event similarity model that underlies the system.

(0 min) Modern applications of survival analysis increasingly involve time-dependent covariates. In healthcare settings, such covariates provide dynamic patient histories that can be used to assess health risks in realtime by tracking the hazard function. Hazard learning is thus particularly useful in healthcare analytics, and the open-source package BoXHED 1.0 provides the first implementation of a gradient boosted hazard estimator that is fully nonparametric. This paper introduces BoXHED 2.0, a quantum leap over BoXHED 1.0 in several ways. Crucially, BoXHED 2.0 can deal with survival data that goes far beyond right-censoring and it also supports recurring events. To our knowledge, this is the only nonparametric machine learning implementation that is able to do so. Another major improvement is that BoXHED 2.0 is orders of magnitude more scalable, due in part to a novel data preprocessing step that sidesteps the need for explicit quadrature when dealing with time-dependent covariates. BoXHED 2.0 supports the use of GPUs and multicore CPUs, and is available from GitHub: www.github.com/BoXHED.

(2 min) Recent history has seen a tremendous growth of work exploring implicit representations of geometry and radiance, popularized through Neural Radiance Fields (NeRF). Such works are fundamentally based on a (implicit) {\em volumetric} representation of occupancy, allowing them to model diverse scene structure including translucent objects and atmospheric obscurants. But because the vast majority of real-world scenes are composed of well-defined surfaces, we introduce a {\em surface} analog of such implicit models called Neural Reflectance Surfaces (NeRS). NeRS learns a neural shape representation of a closed surface that is diffeomorphic to a sphere, guaranteeing water-tight reconstructions. Even more importantly, surface parameterizations allow NeRS to learn (neural) bidirectional surface reflectance functions (BRDFs) that factorize view-dependent appearance into environmental illumination, diffuse color (albedo), and specular "shininess." Finally, rather than illustrating our results on synthetic scenes or controlled in-the-lab capture, we assemble a novel dataset of multi-view images from online marketplaces for selling goods. Such "in-the-wild" multi-view image sets pose a number of challenges, including a small number of views with unknown/rough camera estimates. We demonstrate that surface-based neural reconstructions enable learning from such data, outperforming volumetric neural rendering-based reconstructions. We hope that NeRS serves as a first step toward building scalable, high-quality libraries of real-world shape, materials, and illumination. The project page with code and video visualizations can be found at https://jasonyzhang.com/ners}{jasonyzhang.com/ners.

(2 min) Conventionally, the matrix completion (MC) model aims to recover a matrix from partially observed elements. Accurate recovery necessarily requires a regularization encoding priors of the unknown matrix/signal properly. However, encoding the priors accurately for the complex natural signal is difficult, and even then, the model might not generalize well outside the particular matrix type. This work combines adaptive and implicit low-rank regularization that captures the prior dynamically according to the current recovered matrix. Furthermore, we aim to answer the question: how does adaptive regularization affect implicit regularization? We utilize neural networks to represent Adaptive and Implicit Regularization and named the proposed model \textit{AIR-Net}. Theoretical analyses show that the adaptive part of the AIR-Net enhances implicit regularization. In addition, the adaptive regularizer vanishes at the end, thus can avoid saturation issues. Numerical experiments for various data demonstrate the effectiveness of AIR-Net, especially when the locations of missing elements are not randomly chosen. With complete flexibility to select neural networks for matrix representation, AIR-Net can be extended to solve more general inverse problems.

(2 min) Multi-objective optimization (MOO) aims at finding a set of optimal configurations for a given set of objectives. A recent line of work applies MOO methods to the typical Machine Learning (ML) setting, which becomes multi-objective if a model should optimize more than one objective, for instance in fair machine learning. These works also use Multi-Task Learning (MTL) problems to benchmark MOO algorithms treating each task as independent objective. In this work we show that MTL problems do not resemble the characteristics of MOO problems. In particular, MTL losses are not competing in case of a sufficiently expressive single model. As a consequence, a single model can perform just as well as optimizing all objectives with independent models, rendering MOO inapplicable. We provide evidence with extensive experiments on the widely used Multi-Fashion-MNIST datasets. Our results call for new benchmarks to evaluate MOO algorithms for ML. Our code is available at: https://github.com/ruchtem/moo-mtl.

(2 min) In this paper, we investigate data-driven parameterized modeling of insertion loss for transmission lines with respect to design parameters. We first show that direct application of neural networks can lead to non-physics models with negative insertion loss. To mitigate this problem, we propose two deep learning solutions. One solution is to add a regulation term, which represents the passive condition, to the final loss function to enforce the negative quantity of insertion loss. In the second method, a third-order polynomial expression is defined first, which ensures positiveness, to approximate the insertion loss, then DeepONet neural network structure, which was proposed recently for function and system modeling, was employed to model the coefficients of polynomials. The resulting neural network is applied to predict the coefficients of the polynomial expression. The experimental results on an open-sourced SI/PI database of a PCB design show that both methods can ensure the positiveness for the insertion loss. Furthermore, both methods can achieve similar prediction results, while the polynomial-based DeepONet method is faster than DeepONet based method in training time.

(2 min) Cloth simulation has wide applications in computer animation, garment design, and robot-assisted dressing. This work presents a differentiable cloth simulator whose additional gradient information facilitates cloth-related applications. Our differentiable simulator extends a state-of-the-art cloth simulator based on Projective Dynamics (PD) and with dry frictional contact. We draw inspiration from previous work to propose a fast and novel method for deriving gradients in PD-based cloth simulation with dry frictional contact. Furthermore, we conduct a comprehensive analysis and evaluation of the usefulness of gradients in contact-rich cloth simulation. Finally, we demonstrate the efficacy of our simulator in a number of downstream applications, including system identification, trajectory optimization for assisted dressing, closed-loop control, inverse design, and real-to-sim transfer. We observe a substantial speedup obtained from using our gradient information in solving most of these applications.

(2 min) The distribution of data points is a key component in machine learning. In most cases, one uses min-max normalization to obtain nodes in $[0,1]$ or Z-score normalization for standard normal distributed data. In this paper, we apply transformation ideas in order to design a complete orthonormal system in the $\mathrm{L}_2$ space of functions with the standard normal distribution as integration weight. Subsequently, we are able to apply the explainable ANOVA approximation for this basis and use Z-score transformed data in the method. We demonstrate the applicability of this procedure on the well-known forest fires data set from the UCI machine learning repository. The attribute ranking obtained from the ANOVA approximation provides us with crucial information about which variables in the data set are the most important for the detection of fires.

(2 min) Cheung and Piliouras (2020) recently showed that two variants of the Multiplicative Weights Update method - OMWU and MWU - display opposite convergence properties depending on whether the game is zero-sum or cooperative. Inspired by this work and the recent literature on learning to optimize for single functions, we introduce a new framework for learning last-iterate convergence to Nash Equilibria in games, where the update rule's coefficients (learning rates) along a trajectory are learnt by a reinforcement learning policy that is conditioned on the nature of the game: \textit{the game signature}. We construct the latter using a new decomposition of two-player games into eight components corresponding to commutative projection operators, generalizing and unifying recent game concepts studied in the literature. We compare the performance of various update rules when their coefficients are learnt, and show that the RL policy is able to exploit the game signature across a wide range of game types. In doing so, we introduce CMWU, a new algorithm that extends consensus optimization to the constrained case, has local convergence guarantees for zero-sum bimatrix games, and show that it enjoys competitive performance on both zero-sum games with constant coefficients and across a spectrum of games when its coefficients are learnt.

(2 min) Representation learning from unlabeled data has been of major interest in artificial intelligence research. While self-supervised speech representation learning has been popular in the speech research community, very few works have comprehensively analyzed audio representation learning for non-speech audio tasks. In this paper, we propose a self-supervised audio representation learning method and apply it to a variety of downstream non-speech audio tasks. We combine the well-known wav2vec 2.0 framework, which has shown success in self-supervised learning for speech tasks, with parameter-efficient conformer architectures. On the AudioSet benchmark, we achieve a mean average precision (mAP) score of 0.415, which is a new state-of-the-art on this dataset through audio-only self-supervised learning. Our fine-tuned conformers also surpass or match the performance of previous systems pre-trained in a supervised way on several downstream tasks. We further discuss the important design considerations for both pre-training and fine-tuning.

(2 min) Human behavior modeling deals with learning and understanding of behavior patterns inherent in humans' daily routines. Existing pattern mining techniques either assume human dynamics is strictly periodic, or require the number of modes as input, or do not consider uncertainty in the sensor data. To handle these issues, in this paper, we propose a novel clustering approach for modeling human behavior (named, MTpattern) from time-series data. For mining frequent human behavior patterns effectively, we utilize a three-stage pipeline: (1) represent time series data into sequence of regularly sampled equal-sized unit time intervals for better analysis, (2) a new distance measure scheme is proposed to cluster similar sequences which can handle temporal variation and uncertainty in the data, and (3) exploit an exemplar-based clustering mechanism and fine-tune its parameters to output minimum number of clusters with given permissible distance constraints and without knowing the number of modes present in the data. Then, the average of all sequences in a cluster is considered as a human behavior pattern. Empirical studies on two real-world datasets and a simulated dataset demonstrate the effectiveness of MTpattern w.r.to internal and external measures of clustering.

(2 min) Any-to-any voice conversion technologies convert the vocal timbre of an utterance to any speaker even unseen during training. Although there have been several state-of-the-art any-to-any voice conversion models, they were all based on clean utterances to convert successfully. However, in real-world scenarios, it is difficult to collect clean utterances of a speaker, and they are usually degraded by noises or reverberations. It thus becomes highly desired to understand how these degradations affect voice conversion and build a degradation-robust model. We report in this paper the first comprehensive study on the degradation robustness of any-to-any voice conversion. We show that the performance of state-of-the-art models nowadays was severely hampered given degraded utterances. To this end, we then propose speech enhancement concatenation and denoising training to improve the robustness. In addition to common degradations, we also consider adversarial noises, which alter the model output significantly yet are human-imperceptible. It was shown that both concatenations with off-the-shelf speech enhancement models and denoising training on voice conversion models could improve the robustness, while each of them had pros and cons.

(2 min) Technological progress increasingly envisions the use of robots interacting with people in everyday life. Human-robot collaboration (HRC) is the approach that explores the interaction between a human and a robot, during the completion of an actual physical task. Such interplay is explored both at the cognitive and physical level, by respectively analysing the mutual exchange of information and mechanical power. In HRC works, a cognitive model is typically built, which collects inputs from the environment and from the user, elaborates and translates these into information that can be used by the robot itself. HRC studies progressively employ machine learning algorithms to build the cognitive models and behavioural block that elaborates the acquired external inputs. This is a promising approach still in its early stages and with the potential of significant benefit from the growing field of machine learning. Consequently, this paper proposes a thorough literature review of the use of machine learning techniques in the context of human-robot collaboration. The collection,selection and analysis of the set of 45 key papers, selected from the wide review of the literature on robotics and machine learning, allowed the identification of the current trends in HRC. In particular, a clustering of works based on the type of collaborative tasks, evaluation metrics and cognitive variables modelled is proposed. With these premises, a deep analysis on different families of machine learning algorithms and their properties, along with the sensing modalities used, was carried out. The salient aspects of the analysis are discussed to show trends and suggest possible challenges to tackle in the future research.

(2 min) We study the problem of performance optimization of closed-loop control systems with unmodeled dynamics. Bayesian optimization (BO) has been demonstrated effective for improving closed-loop performance by automatically tuning controller gains or reference setpoints in a model-free manner. However, BO methods have rarely been tested on dynamical systems with unmodeled constraints. In this paper, we propose a violation-aware BO algorithm (VABO) that optimizes closed-loop performance while simultaneously learning constraint-feasible solutions. Unlike classical constrained BO methods which allow an unlimited constraint violations, or safe BO algorithms that are conservative and try to operate with near-zero violations, we allow budgeted constraint violations to improve constraint learning and accelerate optimization. We demonstrate the effectiveness of our proposed VABO method for energy minimization of industrial vapor compression systems.

(2 min) Contrastive self-supervised learning (CSL) has attracted increasing attention for model pre-training via unlabeled data. The resulted CSL models provide instance-discriminative visual features that are uniformly scattered in the feature space. During deployment, the common practice is to directly fine-tune CSL models with cross-entropy, which however may not be the best strategy in practice. Although cross-entropy tends to separate inter-class features, the resulting models still have limited capability for reducing intra-class feature scattering that exists in CSL models. In this paper, we investigate whether applying contrastive learning to fine-tuning would bring further benefits, and analytically find that optimizing the contrastive loss benefits both discriminative representation learning and model optimization during fine-tuning. Inspired by these findings, we propose Contrast-regularized tuning (Core-tuning), a new approach for fine-tuning CSL models. Instead of simply adding the contrastive loss to the objective of fine-tuning, Core-tuning further applies a novel hard pair mining strategy for more effective contrastive fine-tuning, as well as smoothing the decision boundary to better exploit the learned discriminative feature space. Extensive experiments on image classification and semantic segmentation verify the effectiveness of Core-tuning.

(2 min) Neural MMO is a computationally accessible research platform that combines large agent populations, long time horizons, open-ended tasks, and modular game systems. Existing environments feature subsets of these properties, but Neural MMO is the first to combine them all. We present Neural MMO as free and open source software with active support, ongoing development, documentation, and additional training, logging, and visualization tools to help users adapt to this new setting. Initial baselines on the platform demonstrate that agents trained in large populations explore more and learn a progression of skills. We raise other more difficult problems such as many-team cooperation as open research questions which Neural MMO is well-suited to answer. Finally, we discuss current limitations of the platform, potential mitigations, and plans for continued development.

(2 min) Deep generative models have been widely used in several areas of NLP, and various techniques have been proposed to augment them or address their training challenges. In this paper, we propose a simple modification to Variational Autoencoders (VAEs) by using an Isotropic Gaussian Posterior (IGP) that allows for better utilisation of their latent representation space. This model avoids the sub-optimal behavior of VAEs related to inactive dimensions in the representation space. We provide both theoretical analysis, and empirical evidence on various datasets and tasks that show IGP leads to consistent improvement on several quantitative and qualitative grounds, from downstream task performance and sample efficiency to robustness. Additionally, we give insights about the representational properties encouraged by IGP and also show that its gain generalises to image domain as well.

(2 min) In this note we disprove a conjecture of Kuzmin and Warmuth claiming that every family whose VC-dimension is at most d admits an unlabeled compression scheme to a sample of size at most d. We also study the unlabeled compression schemes of the joins of some families and conjecture that these give a larger gap between the VC-dimension and the size of the smallest unlabeled compression scheme for them.

(2 min) Developing an agent in reinforcement learning (RL) that is capable of performing complex control tasks directly from high-dimensional observation such as raw pixels is yet a challenge as efforts are made towards improving sample efficiency and generalization. This paper considers a learning framework for Curiosity Contrastive Forward Dynamics Model (CCFDM) in achieving a more sample-efficient RL based directly on raw pixels. CCFDM incorporates a forward dynamics model (FDM) and performs contrastive learning to train its deep convolutional neural network-based image encoder (IE) to extract conducive spatial and temporal information for achieving a more sample efficiency for RL. In addition, during training, CCFDM provides intrinsic rewards, produced based on FDM prediction error, encourages the curiosity of the RL agent to improve exploration. The diverge and less-repetitive observations provide by both our exploration strategy and data augmentation available in contrastive learning improve not only the sample efficiency but also the generalization. Performance of existing model-free RL methods such as Soft Actor-Critic built on top of CCFDM outperforms prior state-of-the-art pixel-based RL methods on the DeepMind Control Suite benchmark.

(2 min) We study the problem of safe offline reinforcement learning (RL), the goal is to learn a policy that maximizes long-term reward while satisfying safety constraints given only offline data, without further interaction with the environment. This problem is more appealing for real world RL applications, in which data collection is costly or dangerous. Enforcing constraint satisfaction is non-trivial, especially in offline settings, as there is a potential large discrepancy between the policy distribution and the data distribution, causing errors in estimating the value of safety constraints. We show that na\"ive approaches that combine techniques from safe RL and offline RL can only learn sub-optimal solutions. We thus develop a simple yet effective algorithm, Constraints Penalized Q-Learning (CPQ), to solve the problem. Our method admits the use of data generated by mixed behavior policies. We present a theoretical analysis and demonstrate empirically that our approach can learn robustly across a variety of benchmark control tasks, outperforming several baselines.

(2 min) Identifying vulnerabilities in the source code is essential to protect the software systems from cyber security attacks. It, however, is also a challenging step that requires specialized expertise in security and code representation. Inspired by the successful applications of pre-trained programming language (PL) models such as CodeBERT and graph neural networks (GNNs), we propose ReGVD, a general and novel graph neural network-based model for vulnerability detection. In particular, ReGVD views a given source code as a flat sequence of tokens and then examines two effective methods of utilizing unique tokens and indexes respectively to construct a single graph as an input, wherein node features are initialized only by the embedding layer of a pre-trained PL model. Next, ReGVD leverages a practical advantage of residual connection among GNN layers and explores a beneficial mixture of graph-level sum and max poolings to return a graph embedding for the given source code. Experimental results demonstrate that ReGVD outperforms the existing state-of-the-art models and obtain the highest accuracy on the real-world benchmark dataset from CodeXGLUE for vulnerability detection.

(2 min) Learning a generalizable deep model from a few examples in a short time remains a major challenge of machine learning, which has impeded its wide deployment to many scenarios. Recent advances reveal that a properly pre-trained model endows an important property: transferability. A higher transferability of the learned representations indicates a better generalizability across domains of different distributions (domain transferability), or across tasks of different semantics (task transferability). Transferability has become the key to enable data-efficient deep learning, however, existing pre-training methods focus only on the domain transferability while meta-training methods only on the task transferability. This restricts their data-efficiency in downstream scenarios of diverging domains and tasks. A finding of this paper is that even a tight combination of pre-training and meta-training cannot achieve both kinds of transferability. This motivates the proposed Omni-Training framework towards data-efficient deep learning. Our first contribution is Omni-Net, a tri-flow architecture. Besides the joint representation flow, Omni-Net introduces two new parallel flows for pre-training and meta-training, respectively responsible for learning representations of domain transferability and task transferability. Omni-Net coordinates the parallel flows by routing them via the joint-flow, making each gain the other kind of transferability. Our second contribution is Omni-Loss, in which a mean-teacher regularization is imposed to learn generalizable and stabilized representations. Omni-Training is a general framework that accommodates many existing pre-training and meta-training algorithms. A thorough evaluation on cross-task and cross-domain datasets in classification, regression and reinforcement learning problems shows that Omni-Training consistently outperforms the state-of-the-art methods.

(2 min) Classification of large multivariate time series with strong class imbalance is an important task in real-world applications. Standard methods of class weights, oversampling, or parametric data augmentation do not always yield significant improvements for predicting minority classes of interest. Non-parametric data augmentation with Generative Adversarial Networks (GANs) offers a promising solution. We propose Imputation Balanced GAN (IB-GAN), a novel method that joins data augmentation and classification in a one-step process via an imputation-balancing approach. IB-GAN uses imputation and resampling techniques to generate higher quality samples from randomly masked vectors than from white noise, and augments classification through a class-balanced set of real and synthetic samples. Imputation hyperparameter $p_{miss}$ allows for regularization of classifier variability by tuning innovations introduced via generator imputation. IB-GAN is simple to train and model-agnostic, pairing any deep learning classifier with a generator-discriminator duo and resulting in higher accuracy for under-observed classes. Empirical experiments on open-source UCR data and proprietary 90K product dataset show significant performance gains against state-of-the-art parametric and GAN baselines.

(2 min) Random forests on the one hand, and neural networks on the other hand, have met great success in the machine learning community for their predictive performance. Combinations of both have been proposed in the literature, notably leading to the so-called deep forests (DF) (Zhou \& Feng,2019). In this paper, our aim is not to benchmark DF performances but to investigate instead their underlying mechanisms. Additionally, DF architecture can be generally simplified into more simple and computationally efficient shallow forest networks. Despite some instability, the latter may outperform standard predictive tree-based methods. We exhibit a theoretical framework in which a shallow tree network is shown to enhance the performance of classical decision trees. In such a setting, we provide tight theoretical lower and upper bounds on its excess risk. These theoretical results show the interest of tree-network architectures for well-structured data provided that the first layer, acting as a data encoder, is rich enough.

(2 min) Our work aims at efficiently leveraging ambiguous demonstrations for the training of a reinforcement learning (RL) agent. An ambiguous demonstration can usually be interpreted in multiple ways, which severely hinders the RL-Agent from learning stably and efficiently. Since an optimal demonstration may also suffer from being ambiguous, previous works that combine RL and learning from demonstration (RLfD works) may not work well. Inspired by how humans handle such situations, we propose to use self-explanation (an agent generates explanations for itself) to recognize valuable high-level relational features as an interpretation of why a successful trajectory is successful. This way, the agent can provide some guidance for its RL learning. Our main contribution is to propose the Self-Explanation for RL from Demonstrations (SERLfD) framework, which can overcome the limitations of traditional RLfD works. Our experimental results show that an RLfD model can be improved by using our SERLfD framework in terms of training stability and performance.

(2 min) Multi-objective optimization (MOO) is a prevalent challenge for Deep Learning, however, there exists no scalable MOO solution for truly deep neural networks. Prior work either demand optimizing a new network for every point on the Pareto front, or induce a large overhead to the number of trainable parameters by using hyper-networks conditioned on modifiable preferences. In this paper, we propose to condition the network directly on these preferences by augmenting them to the feature space. Furthermore, we ensure a well-spread Pareto front by penalizing the solutions to maintain a small angle to the preference vector. In a series of experiments, we demonstrate that our Pareto fronts achieve state-of-the-art quality despite being computed significantly faster. Furthermore, we showcase the scalability as our method approximates the full Pareto front on the CelebA dataset with an EfficientNet network at a tiny training time overhead of 7% compared to a simple single-objective optimization. We make our code publicly available at https://github.com/ruchtem/cosmos.

(2 min) This paper investigates unsupervised learning of Full-Waveform Inversion (FWI), which has been widely used in geophysics to estimate subsurface velocity maps from seismic data. This problem is mathematically formulated by a second order partial differential equation (PDE), but is hard to solve. Moreover, acquiring velocity map is extremely expensive, making it impractical to scale up a supervised approach to train the mapping from seismic data to velocity maps with convolutional neural networks (CNN). We address these difficulties by integrating PDE and CNN in a loop, thus shifting the paradigm to unsupervised learning that only requires seismic data. In particular, we use finite difference to approximate the forward modeling of PDE as a differentiable operator (from velocity map to seismic data) and model its inversion by CNN (from seismic data to velocity map). Hence, we transform the supervised inversion task into an unsupervised seismic data reconstruction task. We also introduce a new large-scale dataset OpenFWI, to establish a more challenging benchmark for the community. Experiment results show that our model (using seismic data alone) yields comparable accuracy to the supervised counterpart (using both seismic data and velocity map). Furthermore, it outperforms the supervised model when involving more seismic data.

(2 min) We present TWIST, a novel self-supervised representation learning method by classifying large-scale unlabeled datasets in an end-to-end way. We employ a siamese network terminated by a softmax operation to produce twin class distributions of two augmented images. Without supervision, we enforce the class distributions of different augmentations to be consistent. In the meantime, we regularize the class distributions to make them sharp and diverse. Specifically, we minimize the entropy of the distribution for each sample to make the class prediction for each sample assertive and maximize the entropy of the mean distribution to make the predictions of different samples diverse. In this way, TWIST can naturally avoid the trivial solutions without specific designs such as asymmetric network, stop-gradient operation, or momentum encoder. Different from the clustering-based methods which alternate between clustering and learning, our method is a single learning process guided by a unified loss function. As a result, TWIST outperforms state-of-the-art methods on a wide range of tasks, including unsupervised classification, linear classification, semi-supervised learning, transfer learning, and some dense prediction tasks such as detection and segmentation.

(2 min) Equivariance has emerged as a desirable property of representations of objects subject to identity-preserving transformations that constitute a group, such as translations and rotations. However, the expressivity of a representation constrained by group equivariance is still not fully understood. We address this gap by providing a generalization of Cover's Function Counting Theorem that quantifies the number of linearly separable and group-invariant binary dichotomies that can be assigned to equivariant representations of objects. We find that the fraction of separable dichotomies is determined by the dimension of the space that is fixed by the group action. We show how this relation extends to operations such as convolutions, element-wise nonlinearities, and global and local pooling. While other operations do not change the fraction of separable dichotomies, local pooling decreases the fraction, despite being a highly nonlinear operation. Finally, we test our theory on intermediate representations of randomly initialized and fully trained convolutional neural networks and find perfect agreement.

(2 min) Despite the pervasiveness of ordinal labels in supervised learning, it remains common practice in deep learning to treat such problems as categorical classification using the categorical cross entropy loss. Recent methods attempting to address this issue while respecting the ordinal structure of the labels have resorted to converting ordinal regression into a series of extended binary classification subtasks. However, the adoption of such methods remains inconsistent due to theoretical and practical limitations. Here we address these limitations by demonstrating that the subtask probabilities form a Markov chain. We show how to straightforwardly modify neural network architectures to exploit this fact and thereby constrain predictions to be universally rank consistent. We furthermore prove that all rank consistent solutions can be represented within this formulation. Using diverse benchmarks and the real-world application of a specialized recurrent neural network for COVID-19 prognosis, we demonstrate the practical superiority of this method versus the current state-of-the-art. The method is open sourced as user-friendly PyTorch and TensorFlow packages.

(2 min) We introduce a state-of-the-art audio-visual on-screen sound separation system which is capable of learning to separate sounds and associate them with on-screen objects by looking at in-the-wild videos. We identify limitations of previous work on audio-visual on-screen sound separation, including the simplicity and coarse resolution of spatio-temporal attention, and poor convergence of the audio separation model. Our proposed model addresses these issues using cross-modal and self-attention modules that capture audio-visual dependencies at a finer resolution over time, and by unsupervised pre-training of audio separation model. These improvements allow the model to generalize to a much wider set of unseen videos. We also show a robust way to further improve the generalization capability of our models by calibrating the probabilities of our audio-visual on-screen classifier, using only a small amount of in-domain videos labeled for their on-screen presence. For evaluation and semi-supervised training, we collected human annotations of on-screen audio from a large database of in-the-wild videos (YFCC100m). Our results show marked improvements in on-screen separation performance, in more general conditions than previous methods.

(2 min) Power allocation is one of the fundamental problems in wireless networks and a wide variety of algorithms address this problem from different perspectives. A common element among these algorithms is that they rely on an estimation of the channel state, which may be inaccurate on account of hardware defects, noisy feedback systems, and environmental and adversarial disturbances. Therefore, it is essential that the output power allocation of these algorithms is stable with respect to input perturbations, to the extent that the variations in the output are bounded for bounded variations in the input. In this paper, we focus on UWMMSE -- a modern algorithm leveraging graph neural networks --, and illustrate its stability to additive input perturbations of bounded energy through both theoretical analysis and empirical validation.

(2 min) This paper presents a distributionally robust Q-Learning algorithm (DrQ) which leverages Wasserstein ambiguity sets to provide probabilistic out-of-sample safety guarantees during online learning. First, we follow past work by separating the constraint functions from the principal objective to create a hierarchy of machines which estimate the feasible state-action space within the constrained Markov decision process (CMDP). DrQ works within this framework by augmenting constraint costs with tightening offset variables obtained through Wasserstein distributionally robust optimization (DRO). These offset variables correspond to worst-case distributions of modeling error characterized by the TD-errors of the constraint Q-functions. This procedure allows us to safely approach the nominal constraint boundaries with strong probabilistic safety guarantees. Using a case study of safe lithium-ion battery fast charging, we demonstrate dramatic improvements in safety and performance relative to conventional methods.

(2 min) Contemporary natural language processing (NLP) revolves around learning from latent document representations, generated either implicitly by neural language models or explicitly by methods such as doc2vec or similar. One of the key properties of the obtained representations is their dimension. Whilst the commonly adopted dimensions of 256 and 768 offer sufficient performance on many tasks, it is many times unclear whether the default dimension is the most suitable choice for the subsequent downstream learning tasks. Furthermore, representation dimensions are seldom subject to hyperparameter tuning due to computational constraints. The purpose of this paper is to demonstrate that a surprisingly simple and efficient recursive compression procedure can be sufficient to both significantly compress the initial representation, but also potentially improve its performance when considering the task of text classification. Having smaller and less noisy representations is the desired property during deployment, as orders of magnitude smaller models can significantly reduce the computational overload and with it the deployment costs. We propose CoRe, a straightforward, representation learner-agnostic framework suitable for representation compression. The CoRe's performance is showcased and studied on a collection of 17 real-life corpora from biomedical, news, social media, and literary domains. We explored CoRe's behavior when considering contextual and non-contextual document representations, different compression levels, and 9 different compression algorithms. Current results based on more than 100,000 compression experiments indicate that recursive Singular Value Decomposition offers a very good trade-off between the compression efficiency and performance, making CoRe useful in many existing, representation-dependent NLP pipelines.

(2 min) This paper investigates the stochastic distributed nonconvex optimization problem of minimizing a global cost function formed by the summation of $n$ local cost functions. We solve such a problem by involving zeroth-order (ZO) information exchange. In this paper, we propose a ZO distributed primal-dual coordinate method (ZODIAC) to solve the stochastic optimization problem. Agents approximate their own local stochastic ZO oracle along with coordinates with an adaptive smoothing parameter. We show that the proposed algorithm achieves the convergence rate of $\mathcal{O}(\sqrt{p}/\sqrt{T})$ for general nonconvex cost functions. We demonstrate the efficiency of proposed algorithms through a numerical example in comparison with the existing state-of-the-art centralized and distributed ZO algorithms.

(2 min) We consider a mesh-based approach for training a neural network to produce field predictions of solutions to parametric partial differential equations (PDEs). This approach contrasts current approaches for "neural PDE solvers" that employ collocation-based methods to make point-wise predictions of solutions to PDEs. This approach has the advantage of naturally enforcing different boundary conditions as well as ease of invoking well-developed PDE theory -- including analysis of numerical stability and convergence -- to obtain capacity bounds for our proposed neural networks in discretized domains. We explore our mesh-based strategy, called NeuFENet, using a weighted Galerkin loss function based on the Finite Element Method (FEM) on a parametric elliptic PDE. The weighted Galerkin loss (FEM loss) is similar to an energy functional that produces improved solutions, satisfies a priori mesh convergence, and can model Dirichlet and Neumann boundary conditions. We prove theoretically, and illustrate with experiments, convergence results analogous to mesh convergence analysis deployed in finite element solutions to PDEs. These results suggest that a mesh-based neural network approach serves as a promising approach for solving parametric PDEs with theoretical bounds.

(0 min) We compare classification and regression tasks in an overparameterized linear model with Gaussian features. On the one hand, we show that with sufficient overparameterization all training points are support vectors: solutions obtained by least-squares minimum-norm interpolation, typically used for regression, are identical to those produced by the hard-margin support vector machine (SVM) that minimizes the hinge loss, typically used for training classifiers. On the other hand, we show that there exist regimes where these interpolating solutions generalize well when evaluated by the 0-1 test loss function, but do not generalize if evaluated by the square loss function, i.e. they approach the null risk. Our results demonstrate the very different roles and properties of loss functions used at the training phase (optimization) and the testing phase (generalization).

(0 min) Attack graphs (AG) are used to assess pathways availed by cyber adversaries to penetrate a network. State-of-the-art approaches for AG generation focus mostly on deriving dependencies between system vulnerabilities based on network scans and expert knowledge. In real-world operations however, it is costly and ineffective to rely on constant vulnerability scanning and expert-crafted AGs. We propose to automatically learn AGs based on actions observed through intrusion alerts, without prior expert knowledge. Specifically, we develop an unsupervised sequence learning system, SAGE, that leverages the temporal and probabilistic dependence between alerts in a suffix-based probabilistic deterministic finite automaton (S-PDFA) -- a model that accentuates infrequent severe alerts and summarizes paths leading to them. AGs are then derived from the S-PDFA on a per-objective, per-victim basis. Tested with intrusion alerts collected through Collegiate Penetration Testing Competition, SAGE compresses over 330k alerts into 93 AGs. These AGs reflect the strategies used by the participating teams. The AGs are succinct, interpretable, and capture behavioral dynamics, e.g., that attackers will often follow shorter paths to re-exploit objectives.

(0 min) We present a novel generative modeling method called diffusion normalizing flow based on stochastic differential equations (SDEs). The algorithm consists of two neural SDEs: a forward SDE that gradually adds noise to the data to transform the data into Gaussian random noise, and a backward SDE that gradually removes the noise to sample from the data distribution. By jointly training the two neural SDEs to minimize a common cost function that quantifies the difference between the two, the backward SDE converges to a diffusion process the starts with a Gaussian distribution and ends with the desired data distribution. Our method is closely related to normalizing flow and diffusion probabilistic models and can be viewed as a combination of the two. Compared with normalizing flow, diffusion normalizing flow is able to learn distributions with sharp boundaries. Compared with diffusion probabilistic models, diffusion normalizing flow requires fewer discretization steps and thus has better sampling efficiency. Our algorithm demonstrates competitive performance in both high-dimension data density estimation and image generation tasks.

(0 min) The 2021 Drexel Society of Artificial Intelligence Research Conference highlights papers focused on a broad set of papers in machine learning. This was our organizations' first annual conference. It was conducted virtually via Zoom. The highlights are currently posted on YouTube.

(0 min) Single-image generative adversarial networks learn from the internal distribution of a single training example to generate variations of it, removing the need of a large dataset. In this paper we introduce SpecSinGAN, an unconditional generative architecture that takes a single one-shot sound effect (e.g., a footstep; a character jump) and produces novel variations of it, as if they were different takes from the same recording session. We explore the use of multi-channel spectrograms to train the model on the various layers that comprise a single sound effect. A listening study comparing our model to real recordings and to digital signal processing procedural audio models in terms of sound plausibility and variation revealed that SpecSinGAN is more plausible and varied than the procedural audio models considered, when using multi-channel spectrograms. Sound examples can be found at the project website: https://www.adrianbarahonarios.com/specsingan/

(0 min) Modern Internet of Things solutions are used in a variety of different areas, ranging from connected vehicles and healthcare to industrial applications. They rely on a large amount of interconnected sensors, which can lead to both technical and economical challenges. Virtual sensing techniques aim to reduce the number of physical sensors in a system by using data from available measurements to estimate additional unknown quantities of interest. Successful model-based solutions include Kalman filters or the combination of finite element models and modal analysis, while many data-driven methods rely on machine learning algorithms. The presented hybrid virtual sensing approach combines Long Short-Term Memory networks with frequency response function models in order to estimate the behavior of non-linear dynamic systems with multiple input and output channels. Network training and prediction make use of short signal subsequences, which are later recombined by applying a windowing technique. The frequency response function model acts as a baseline estimate which perfectly captures linear dynamic systems and is augmented by the non-linear Long Short-Term Memory network following two different hybrid modeling strategies. The approach is tested using a non-linear experimental dataset, which results from measurements of a three-component servo-hydraulic fatigue test bench. A variety of metrics in time and frequency domains, as well as fatigue strength under variable amplitudes are used to evaluate the approximation quality of the proposed method. In addition to virtual sensing, the algorithm is also applied to a forward prediction task. Synthetic data are used in a separate study to estimate the prediction quality on datasets of different size.

(0 min) With the growing interest in the machine learning community to solve real-world problems, it has become crucial to uncover the hidden reasoning behind their decisions by focusing on the fairness and auditing the predictions made by these black-box models. In this paper, we propose a novel method to address two key issues: (a) Can we simultaneously learn fair disentangled representations while ensuring the utility of the learned representation for downstream tasks, and (b)Can we provide theoretical insights into when the proposed approach will be both fair and accurate. To address the former, we propose the method FRIED, Fair Representation learning using Interpolation Enabled Disentanglement. In our architecture, by imposing a critic-based adversarial framework, we enforce the interpolated points in the latent space to be more realistic. This helps in capturing the data manifold effectively and enhances the utility of the learned representation for downstream prediction tasks. We address the latter question by developing a theory on fairness-accuracy trade-offs using classifier-based conditional mutual information estimation. We demonstrate the effectiveness of FRIED on datasets of different modalities - tabular, text, and image datasets. We observe that the representations learned by FRIED are overall fairer in comparison to existing baselines and also accurate for downstream prediction tasks. Additionally, we evaluate FRIED on a real-world healthcare claims dataset where we conduct an expert aided model auditing study providing useful insights into opioid ad-diction patterns.

(0 min) Given the prevalence of large-scale graphs in real-world applications, the storage and time for training neural models have raised increasing concerns. To alleviate the concerns, we propose and study the problem of graph condensation for graph neural networks (GNNs). Specifically, we aim to condense the large, original graph into a small, synthetic and highly-informative graph, such that GNNs trained on the small graph and large graph have comparable performance. We approach the condensation problem by imitating the GNN training trajectory on the original graph through the optimization of a gradient matching loss and design a strategy to condense node futures and structural information simultaneously. Extensive experiments have demonstrated the effectiveness of the proposed framework in condensing different graph datasets into informative smaller graphs. In particular, we are able to approximate the original test accuracy by 95.3% on Reddit, 99.8% on Flickr and 99.0% on Citeseer, while reducing their graph size by more than 99.9%, and the condensed graphs can be used to train various GNN architectures.

(0 min) In attempts to develop sample-efficient algorithms, researcher have explored myriad mechanisms for collecting and exploiting feature feedback, auxiliary annotations provided for training (but not test) instances that highlight salient evidence. Examples include bounding boxes around objects and salient spans in text. Despite its intuitive appeal, feature feedback has not delivered significant gains in practical problems as assessed on iid holdout sets. However, recent works on counterfactually augmented data suggest an alternative benefit of supplemental annotations: lessening sensitivity to spurious patterns and consequently delivering gains in out-of-domain evaluations. Inspired by these findings, we hypothesize that while the numerous existing methods for incorporating feature feedback have delivered negligible in-sample gains, they may nevertheless generalize better out-of-domain. In experiments addressing sentiment analysis, we show that feature feedback methods perform significantly better on various natural out-of-domain datasets even absent differences on in-domain evaluation. By contrast, on natural language inference tasks, performance remains comparable. Finally, we compare those tasks where feature feedback does (and does not) help.

(0 min) One of the lessons from the COVID-19 pandemic is the importance of social distancing, even in challenging circumstances such as pre-hurricane evacuation. To explore the implications of integrating social distancing with evacuation operations, we describe this evacuation process as a Capacitated Vehicle Routing Problem (CVRP) and solve it using a DNN (Deep Neural Network)-based solution (Deep Reinforcement Learning) and a non-DNN solution (Sweep Algorithm). A central question is whether Deep Reinforcement Learning provides sufficient extra routing efficiency to accommodate increased social distancing in a time-constrained evacuation operation. We found that, in comparison to the Sweep Algorithm, Deep Reinforcement Learning can provide decision-makers with more efficient routing. However, the evacuation time saved by Deep Reinforcement Learning does not come close to compensating for the extra time required for social distancing, and its advantage disappears as the emergency vehicle capacity approaches the number of people per household.

(0 min) We propose an adaptive (stochastic) gradient perturbation method for differentially private empirical risk minimization. At each iteration, the random noise added to the gradient is optimally adapted to the stepsize; we name this process adaptive differentially private (ADP) learning. Given the same privacy budget, we prove that the ADP method considerably improves the utility guarantee compared to the standard differentially private method in which vanilla random noise is added. Our method is particularly useful for gradient-based algorithms with time-varying learning rates, including variants of AdaGrad (Duchi et al., 2011). We provide extensive numerical experiments to demonstrate the effectiveness of the proposed adaptive differentially private algorithm.

(0 min) While unbiased machine learning models are essential for many applications, bias is a human-defined concept that can vary across tasks. Given only input-label pairs, algorithms may lack sufficient information to distinguish stable (causal) features from unstable (spurious) features. However, related tasks often share similar biases -- an observation we may leverage to develop stable classifiers in the transfer setting. In this work, we explicitly inform the target classifier about unstable features in the source tasks. Specifically, we derive a representation that encodes the unstable features by contrasting different data environments in the source task. We achieve robustness by clustering data of the target task according to this representation and minimizing the worst-case risk across these clusters. We evaluate our method on both text and image classifications. Empirical results demonstrate that our algorithm is able to maintain robustness on the target task, outperforming the best baseline by 22.9% in absolute accuracy across 12 transfer settings. Our code is available at https://github.com/YujiaBao/Tofu.

(2 min) Knowledge tracing (KT) has recently been an active research area of computational pedagogy. The task is to model students' mastery level of knowledge concepts based on their responses to the questions in the past, as well as predict the probabilities that they correctly answer subsequent questions in the future. KT tasks were historically solved using statistical modeling methods such as Bayesian inference and factor analysis, but recent advances in deep learning have led to the successive proposals that leverage deep neural networks, including long short-term memory networks, memory-augmented networks and self-attention networks. While those deep models demonstrate superior performance over the traditional approaches, they all neglect the explicit modeling of the learning curve theory, which generally says that more practice on the same knowledge concept enhances one's mastery level of the concept. Based on this theory, we propose a Convolution-Augmented Knowledge Tracing (CAKT) model in this paper. The model employs three-dimensional convolutional neural networks to explicitly learn a student's recent experience on applying the same knowledge concept with that in the next question, and fuses the learnt feature with the feature representing her overall latent knowledge state obtained using a classic LSTM network. The fused feature is then fed into a second LSTM network to predict the student's response to the next question. Experimental results show that CAKT achieves the new state-of-the-art performance in predicting students' responses compared with existing models. We also conduct extensive sensitivity analysis and ablation study to show the stability of the results and justify the particular architecture of CAKT, respectively.

(2 min) The main challenge for domain generalization (DG) is to overcome the potential distributional shift between multiple training domains and unseen test domains. One popular class of DG algorithms aims to learn representations that have an invariant causal relation across the training domains. However, certain features, called \emph{pseudo-invariant features}, may be invariant in the training domain but not the test domain and can substantially decreases the performance of existing algorithms. To address this issue, we propose a novel algorithm, called Invariant Information Bottleneck (IIB), that learns a minimally sufficient representation that is invariant across training and testing domains. By minimizing the mutual information between the representation and inputs, IIB alleviates its reliance on pseudo-invariant features, which is desirable for DG. To verify the effectiveness of the IIB principle, we conduct extensive experiments on large-scale DG benchmarks. The results show that IIB outperforms invariant learning baseline (e.g. IRM) by an average of 2.8\% and 3.8\% accuracy over two evaluation metrics.

(2 min) We describe a simple pre-training approach for point clouds. It works in three steps: 1. Mask all points occluded in a camera view; 2. Learn an encoder-decoder model to reconstruct the occluded points; 3. Use the encoder weights as initialisation for downstream point cloud tasks. We find that even when we construct a single pre-training dataset (from ModelNet40), this pre-training method improves accuracy across different datasets and encoders, on a wide range of downstream tasks. Specifically, we show that our method outperforms previous pre-training methods in object classification, and both part-based and semantic segmentation tasks. We study the pre-trained features and find that they lead to wide downstream minima, have high transformation invariance, and have activations that are highly correlated with part labels. Code and data are available at: https://github.com/hansen7/OcCo

(2 min) Autonomous cyber-physical systems (CPS) can improve safety and efficiency for safety-critical applications, but require rigorous testing before deployment. The complexity of these systems often precludes the use of formal verification and real-world testing can be too dangerous during development. Therefore, simulation-based techniques have been developed that treat the system under test as a black box operating in a simulated environment. Safety validation tasks include finding disturbances in the environment that cause the system to fail (falsification), finding the most-likely failure, and estimating the probability that the system fails. Motivated by the prevalence of safety-critical artificial intelligence, this work provides a survey of state-of-the-art safety validation techniques for CPS with a focus on applied algorithms and their modifications for the safety validation problem. We present and discuss algorithms in the domains of optimization, path planning, reinforcement learning, and importance sampling. Problem decomposition techniques are presented to help scale algorithms to large state spaces, which are common for CPS. A brief overview of safety-critical applications is given, including autonomous vehicles and aircraft collision avoidance systems. Finally, we present a survey of existing academic and commercially available safety validation tools.

(2 min) White-box Adversarial Example (AE) attacks towards Deep Neural Networks (DNNs) have a more powerful destructive capacity than black-box AE attacks in the fields of AE strategies. However, almost all the white-box approaches lack interpretation from the point of view of DNNs. That is, adversaries did not investigate the attacks from the perspective of interpretable features, and few of these approaches considered what features the DNN actually learns. In this paper, we propose an interpretable white-box AE attack approach, DI-AA, which explores the application of the interpretable approach of the deep Taylor decomposition in the selection of the most contributing features and adopts the Lagrangian relaxation optimization of the logit output and L_p norm to further decrease the perturbation. We compare DI-AA with six baseline attacks (including the state-of-the-art attack AutoAttack) on three datasets. Experimental results reveal that our proposed approach can 1) attack non-robust models with comparatively low perturbation, where the perturbation is closer to or lower than the AutoAttack approach; 2) break the TRADES adversarial training models with the highest success rate; 3) the generated AE can reduce the robust accuracy of the robust black-box models by 16% to 31% in the black-box transfer attack.

(2 min) Style transfer is the task of rewriting an input sentence into a target style while approximately preserving its content. While most prior literature assumes access to large style-labelled corpora, recent work (Riley et al. 2021) has attempted "few-shot" style transfer using only 3-10 sentences at inference for extracting the target style. In this work we consider one such low resource setting where no datasets are available: style transfer for Indian languages. We find that existing few-shot methods perform this task poorly, with a strong tendency to copy inputs verbatim. We push the state-of-the-art for few-shot style transfer with a new method modeling the stylistic difference between paraphrases. When compared to prior work using automatic and human evaluations, our model achieves 2-3x better performance and output diversity in formality transfer and code-mixing addition across five Indian languages. Moreover, our method is better able to control the amount of style transfer using an input scalar knob. We report promising qualitative results for several attribute transfer directions, including sentiment transfer, text simplification, gender neutralization and text anonymization, all without retraining the model. Finally we found model evaluation to be difficult due to the lack of evaluation datasets and metrics for Indian languages. To facilitate further research in formality transfer for Indic languages, we crowdsource annotations for 4000 sentence pairs in four languages, and use this dataset to design our automatic evaluation suite.

(2 min) Network representation learning (NRL) advances the conventional graph mining of social networks, knowledge graphs, and complex biomedical and physics information networks. Over dozens of network representation learning algorithms have been reported in the literature. Most of them focus on learning node embeddings for homogeneous networks, but they differ in the specific encoding schemes and specific types of node semantics captured and used for learning node embedding. This survey paper reviews the design principles and the different node embedding techniques for network representation learning over homogeneous networks. To facilitate the comparison of different node embedding algorithms, we introduce a unified reference framework to divide and generalize the node embedding learning process on a given network into preprocessing steps, node feature extraction steps and node embedding model training for a NRL task such as link prediction and node clustering. With this unifying reference framework, we highlight the representative methods, models, and techniques used at different stages of the node embedding model learning process. This survey not only helps researchers and practitioners to gain an in-depth understanding of different network representation learning techniques but also provides practical guidelines for designing and developing the next generation of network representation learning algorithms and systems.

(2 min) This paper presents a new neural architecture that combines a modulated Hebbian network (MOHN) with DQN, which we call modulated Hebbian plus Q network architecture (MOHQA). The hypothesis is that such a combination allows MOHQA to solve difficult partially observable Markov decision process (POMDP) problems which impair temporal difference (TD)-based RL algorithms such as DQN, as the TD error cannot be easily derived from observations. The key idea is to use a Hebbian network with bio-inspired neural traces in order to bridge temporal delays between actions and rewards when confounding observations and sparse rewards result in inaccurate TD errors. In MOHQA, DQN learns low level features and control, while the MOHN contributes to the high-level decisions by associating rewards with past states and actions. Thus the proposed architecture combines two modules with significantly different learning algorithms, a Hebbian associative network and a classical DQN pipeline, exploiting the advantages of both. Simulations on a set of POMDPs and on the MALMO environment show that the proposed algorithm improved DQN's results and even outperformed control tests with A2C, QRDQN+LSTM and REINFORCE algorithms on some POMDPs with confounding stimuli and sparse rewards.

(3 min) Recent works on sparse neural networks have demonstrated the possibility to train a sparse subnetwork independently from scratch, to match the performance of its corresponding dense network. However, identifying such sparse subnetworks (winning tickets) either involves a costly iterative train-prune-retrain process (e.g., Lottery Ticket Hypothesis) or an over-extended training time (e.g., Dynamic Sparse Training). In this work, we draw a unique connection between sparse neural network training and the deep ensembling technique, yielding a novel ensemble learning framework called FreeTickets. Instead of starting from a dense network, FreeTickets randomly initializes a sparse subnetwork and then trains the subnetwork while dynamically adjusting its sparse mask, resulting in many diverse sparse subnetworks throughout the training process. FreeTickets is defined as the ensemble of these sparse subnetworks freely obtained during this one-pass, sparse-to-sparse training, which uses only a fraction of the computational resources required by the vanilla dense training. Moreover, despite being an ensemble of models, FreeTickets has even fewer parameters and training FLOPs compared to a single dense model: this seemingly counter-intuitive outcome is due to the high sparsity of each subnetwork. FreeTickets is observed to demonstrate a significant all-round improvement compared to standard dense baselines, in prediction accuracy, uncertainty estimation, robustness, and efficiency. FreeTickets easily outperforms the naive deep ensemble with ResNet50 on ImageNet using only a quarter of the training FLOPs required by the latter. Our results provide insights into the strength of sparse neural networks and suggest that the benefits of sparsity go way beyond the usually expected inference efficiency.

(2 min) Nowadays, using vibration data in conjunction with pattern recognition methods is one of the most common fault detection strategies for structures. However, their performances depend on the features extracted from vibration data, the features selected to train the classifier, and the classifier used for pattern recognition. Deep learning facilitates the fault detection procedure by automating the feature extraction and selection, and classification procedure. Though, deep learning approaches have challenges in designing its structure and tuning its hyperparameters, which may result in a low generalization capability. Therefore, this study proposes an ensemble deep learning framework based on a convolutional neural network (CNN) and Dempster-Shafer theory (DST), called CNN-DST. In this framework, several CNNs with the proposed structure are first trained, and then, the outputs of the CNNs selected by the proposed technique are combined by using an improved DST-based method. To validate the proposed CNN-DST framework, it is applied to an experimental dataset created by the broadband vibrational responses of polycrystalline Nickel alloy first-stage turbine blades with different types and severities of damage. Through statistical analysis, it is shown that the proposed CNN-DST framework classifies the turbine blades with an average prediction accuracy of 97.19%. The proposed CNN-DST framework is benchmarked with other state-of-the-art classification methods, demonstrating its high performance. The robustness of the proposed CNN-DST framework with respect to measurement noise is investigated, showing its high noise-resistance. Further, bandwidth analysis reveals that most of the required information for detecting faulty samples is available in a small frequency range.

(3 min) Messenger RNA-based medicines hold immense potential, as evidenced by their rapid deployment as COVID-19 vaccines. However, worldwide distribution of mRNA molecules has been limited by their thermostability, which is fundamentally limited by the intrinsic instability of RNA molecules to a chemical degradation reaction called in-line hydrolysis. Predicting the degradation of an RNA molecule is a key task in designing more stable RNA-based therapeutics. Here, we describe a crowdsourced machine learning competition ("Stanford OpenVaccine") on Kaggle, involving single-nucleotide resolution measurements on 6043 102-130-nucleotide diverse RNA constructs that were themselves solicited through crowdsourcing on the RNA design platform Eterna. The entire experiment was completed in less than 6 months. Winning models demonstrated test set errors that were better by 50% than the previous state-of-the-art DegScore model. Furthermore, these models generalized to blindly predicting orthogonal degradation data on much longer mRNA molecules (504-1588 nucleotides) with improved accuracy over DegScore and other models. Top teams integrated natural language processing architectures and data augmentation techniques with predictions from previous dynamic programming models for RNA secondary structure. These results indicate that such models are capable of representing in-line hydrolysis with excellent accuracy, supporting their use for designing stabilized messenger RNAs. The integration of two crowdsourcing platforms, one for data set creation and another for machine learning, may be fruitful for other urgent problems that demand scientific discovery on rapid timescales.

(2 min) With proliferation of deep learning (DL) applications in diverse domains, vulnerability of DL models to adversarial attacks has become an increasingly interesting research topic in the domains of Computer Vision (CV) and Natural Language Processing (NLP). DL has also been widely adopted to diverse PHM applications, where data are primarily time-series sensor measurements. While those advanced DL algorithms/models have resulted in an improved PHM algorithms' performance, the vulnerability of those PHM algorithms to adversarial attacks has not drawn much attention in the PHM community. In this paper we attempt to explore the vulnerability of PHM algorithms. More specifically, we investigate the strategies of attacking PHM algorithms by considering several unique characteristics associated with time-series sensor measurements data. We use two real-world PHM applications as examples to validate our attack strategies and to demonstrate that PHM algorithms indeed are vulnerable to adversarial attacks.

(2 min) Dense retrieval (DR) methods conduct text retrieval by first encoding texts in the embedding space and then matching them by nearest neighbor search. This requires strong locality properties from the representation space, i.e, the close allocations of each small group of relevant texts, which are hard to generalize to domains without sufficient training data. In this paper, we aim to improve the generalization ability of DR models from source training domains with rich supervision signals to target domains without any relevant labels, in the zero-shot setting. To achieve that, we propose Momentum adversarial Domain Invariant Representation learning (MoDIR), which introduces a momentum method in the DR training process to train a domain classifier distinguishing source versus target, and then adversarially updates the DR encoder to learn domain invariant representations. Our experiments show that MoDIR robustly outperforms its baselines on 10+ ranking datasets from the BEIR benchmark in the zero-shot setup, with more than 10% relative gains on datasets with enough sensitivity for DR models' evaluation. Source code of this paper will be released.

(2 min) A classical computer works with ones and zeros, whereas a quantum computer uses ones, zeros, and superpositions of ones and zeros, which enables quantum computers to perform a vast number of calculations simultaneously compared to classical computers. In a cloud-supported cyber-physical system environment, running a machine learning application in quantum computers is often difficult, due to the existing limitations of the current quantum devices. However, with the combination of quantum-classical neural networks (NN), complex and high-dimensional features can be extracted by the classical NN to a reduced but more informative feature space to be processed by the existing quantum computers. In this study, we develop a hybrid quantum-classical NN to detect an amplitude shift cyber-attack on an in-vehicle control area network (CAN) dataset. We show that using the hybrid quantum classical NN, it is possible to achieve an attack detection accuracy of 94%, which is higher than a Long short-term memory (LSTM) NN (87%) or quantum NN alone (62%)

(2 min) Efficient collaboration between collaborative machine learning and wireless communication technology, forming a Federated Edge Learning (FEEL), has spawned a series of next-generation intelligent applications. However, due to the openness of network connections, the FEEL framework generally involves hundreds of remote devices (or clients), resulting in expensive communication costs, which is not friendly to resource-constrained FEEL. To address this issue, we propose a distributed approximate Newton-type algorithm with fast convergence speed to alleviate the problem of FEEL resource (in terms of communication resources) constraints. Specifically, the proposed algorithm is improved based on distributed L-BFGS algorithm and allows each client to approximate the high-cost Hessian matrix by computing the low-cost Fisher matrix in a distributed manner to find a "better" descent direction, thereby speeding up convergence. Second, we prove that the proposed algorithm has linear convergence in strongly convex and non-convex cases and analyze its computational and communication complexity. Similarly, due to the heterogeneity of the connected remote devices, FEEL faces the challenge of heterogeneous data and non-IID (Independent and Identically Distributed) data. To this end, we design a simple but elegant training scheme, namely FedOVA, to solve the heterogeneous statistical challenge brought by heterogeneous data. In this way, FedOVA first decomposes a multi-class classification problem into more straightforward binary classification problems and then combines their respective outputs using ensemble learning. In particular, the scheme can be well integrated with our communication efficient algorithm to serve FEEL. Numerical results verify the effectiveness and superiority of the proposed algorithm.

(2 min) Automated audio captioning (AAC) is the task of automatically generating textual descriptions for general audio signals. A captioning system has to identify various information from the input signal and express it with natural language. Existing works mainly focus on investigating new methods and try to improve their performance measured on existing datasets. Having attracted attention only recently, very few works on AAC study the performance of existing pre-trained audio and natural language processing resources. In this paper, we evaluate the performance of off-the-shelf models with a Transformer-based captioning approach. We utilize the freely available Clotho dataset to compare four different pre-trained machine listening models, four word embedding models, and their combinations in many different settings. Our evaluation suggests that YAMNet combined with BERT embeddings produces the best captions. Moreover, in general, fine-tuning pre-trained word embeddings can lead to better performance. Finally, we show that sequences of audio embeddings can be processed using a Transformer encoder to produce higher-quality captions.

(2 min) We propose a novel method for creating adversarial examples. Instead of perturbing pixels, we use an encoder-decoder representation of the input image and perturb intermediate layers in the decoder. This changes the high-level features provided by the generative model. Therefore, our perturbation possesses semantic meaning, such as a longer beak or green tints. We formulate this task as an optimization problem by minimizing the Wasserstein distance between the adversarial and initial images under a misclassification constraint. We employ the projected gradient method with a simple inexact projection. Due to the projection, all iterations are feasible, and our method always generates adversarial images. We perform numerical experiments on the MNIST and ImageNet datasets in both targeted and untargeted settings. We demonstrate that our adversarial images are much less vulnerable to steganographic defence techniques than pixel-based attacks. Moreover, we show that our method modifies key features such as edges and that defence techniques based on adversarial training are vulnerable to our attacks.

(2 min) Due to the development of graph neural networks, graph-based representation learning methods have made great progress in recommender systems. However, data sparsity is still a challenging problem that most graph-based recommendation methods are confronted with. Recent works try to address this problem by utilizing side information. In this paper, we exploit the relevant and easily accessible textual information by advanced natural language processing (NLP) models and propose a light RGCN-based (RGCN, relational graph convolutional network) collaborative filtering method based on heterogeneous graphs. Specifically, to incorporate rich textual knowledge, we utilize a pre-trained NLP model to initialize the embeddings of text nodes. Afterward, by performing a simplified RGCN-based node information propagation on the constructed heterogeneous graph, the embeddings of users and items can be adjusted with textual knowledge, which effectively alleviates the negative effects of data sparsity. Moreover, the matching function used by most graph-based representation learning methods is the inner product, which is not appropriate for the obtained embeddings that contain complex semantics. We design a predictive network that combines graph-based representation learning with neural matching function learning, and demonstrate that this architecture can bring a significant performance improvement. Extensive experiments are conducted on three publicly available datasets, and the results verify the superior performance of our method over several baselines.

(2 min) Methods for model explainability have become increasingly critical for testing the fairness and soundness of deep learning. A number of explainability techniques have been developed which use a set of examples to represent a human-interpretable concept in a model's activations. In this work we show that these explainability methods can suffer the same vulnerability to adversarial attacks as the models they are meant to analyze. We demonstrate this phenomenon on two well-known concept-based approaches to the explainability of deep learning models: TCAV and faceted feature visualization. We show that by carefully perturbing the examples of the concept that is being investigated, we can radically change the output of the interpretability method, e.g. showing that stripes are not an important factor in identifying images of a zebra. Our work highlights the fact that in safety-critical applications, there is need for security around not only the machine learning pipeline but also the model interpretation process.

(2 min) Drone-based vehicle detection aims at finding the vehicle locations and categories in an aerial image. It empowers smart city traffic management and disaster rescue. Researchers have made mount of efforts in this area and achieved considerable progress. Nevertheless, it is still a challenge when the objects are hard to distinguish, especially in low light conditions. To tackle this problem, we construct a large-scale drone-based RGB-Infrared vehicle detection dataset, termed DroneVehicle. Our DroneVehicle collects 28, 439 RGB-Infrared image pairs, covering urban roads, residential areas, parking lots, and other scenarios from day to night. Due to the great gap between RGB and infrared images, cross-modal images provide both effective information and redundant information. To address this dilemma, we further propose an uncertainty-aware cross-modality vehicle detection (UA-CMDet) framework to extract complementary information from cross-modal images, which can significantly improve the detection performance in low light conditions. An uncertainty-aware module (UAM) is designed to quantify the uncertainty weights of each modality, which is calculated by the cross-modal Intersection over Union (IoU) and the RGB illumination value. Furthermore, we design an illumination-aware cross-modal non-maximum suppression algorithm to better integrate the modal-specific information in the inference phase. Extensive experiments on the DroneVehicle dataset demonstrate the flexibility and effectiveness of the proposed method for crossmodality vehicle detection. The dataset can be download from https://github.com/VisDrone/DroneVehicle.

(2 min) The goal of Word Sense Disambiguation (WSD) is to identify the sense of a polysemous word in a specific context. Deep-learning techniques using BERT have achieved very promising results in the field and different methods have been proposed to integrate structured knowledge to enhance performance. At the same time, an increasing number of data augmentation techniques have been proven to be useful for NLP tasks. Building upon previous works leveraging BERT and WordNet knowledge, we explore different data augmentation techniques on context-gloss pairs to improve the performance of WSD. In our experiment, we show that both sentence-level and word-level augmentation methods are effective strategies for WSD. Also, we find out that performance can be improved by adding hypernyms' glosses obtained from a lexical knowledge base. We compare and analyze different context-gloss augmentation techniques, and the results show that applying back translation on gloss performs the best.

(2 min) Most prior approaches to offline reinforcement learning (RL) utilize \textit{behavior regularization}, typically augmenting existing off-policy actor critic algorithms with a penalty measuring divergence between the policy and the offline data. However, these approaches lack guaranteed performance improvement over the behavior policy. In this work, we start from the performance difference between the learned policy and the behavior policy, we derive a new policy learning objective that can be used in the offline setting, which corresponds to the advantage function value of the behavior policy, multiplying by a state-marginal density ratio. We propose a practical way to compute the density ratio and demonstrate its equivalence to a state-dependent behavior regularization. Unlike state-independent regularization used in prior approaches, this \textit{soft} regularization allows more freedom of policy deviation at high confidence states, leading to better performance and stability. We thus term our resulting algorithm Soft Behavior-regularized Actor Critic (SBAC). Our experimental results show that SBAC matches or outperforms the state-of-the-art on a set of continuous control locomotion and manipulation tasks.

(2 min) We present a data-driven analysis of MOCK, $\Delta$-MOCK, and MOCLE. These are three closely related approaches that use multi-objective optimization for crisp clustering. More specifically, based on a collection of 12 datasets presenting different proprieties, we investigate the performance of MOCLE and MOCK compared to the recently proposed $\Delta$-MOCK. Besides performing a quantitative analysis identifying which method presents a good/poor performance with respect to another, we also conduct a more detailed analysis on why such a behavior happened. Indeed, the results of our analysis provide useful insights into the strengths and weaknesses of the methods investigated.

(2 min) Multi-agent reinforcement learning often suffers from the exponentially larger action space caused by a large number of agents. In this paper, we propose a novel value decomposition framework HAVEN based on hierarchical reinforcement learning for the fully cooperative multi-agent problems. In order to address instabilities that arise from the concurrent optimization of high-level and low-level policies and another concurrent optimization of agents, we introduce the dual coordination mechanism of inter-layer strategies and inter-agent strategies. HAVEN does not require domain knowledge and pretraining at all, and can be applied to any value decomposition variants. Our method is demonstrated to achieve superior results to many baselines on StarCraft II micromanagement tasks and offers an efficient solution to multi-agent hierarchical reinforcement learning in fully cooperative scenarios.

(2 min) Despite the enormous success of graph neural networks (GNNs), most existing GNNs can only be applicable to undirected graphs where relationships among connected nodes are two-way symmetric (i.e., information can be passed back and forth). However, there is a vast amount of applications where the information flow is asymmetric, leading to directed graphs where information can only be passed in one direction. For example, a directed edge indicates that the information can only be conveyed forwardly from the start node to the end node, but not backwardly. To accommodate such an asymmetric structure of directed graphs within the framework of GNNs, we propose a simple yet remarkably effective framework for directed graph analysis to incorporate such one-way information passing. We define an incoming embedding and an outgoing embedding for each node to model its sending and receiving features respectively. We further develop two steps in our directed GNN model with the first one to aggregate/update the incoming features of nodes and the second one to aggregate/update the outgoing features. By imposing the two roles for each node, the likelihood of a directed edge can be calculated based on the outgoing embedding of the start node and the incoming embedding of the end node. The log-likelihood of all edges plays a natural role of regularization for the proposed model, which can alleviate the over-smoothing problem of the deep GNNs. Extensive experiments on multiple real-world directed graphs demonstrate outstanding performances of the proposed model in both node-level and graph-level tasks.

(2 min) We consider the problem of finding the best memoryless stochastic policy for an infinite-horizon partially observable Markov decision process (POMDP) with finite state and action spaces with respect to either the discounted or mean reward criterion. We show that the (discounted) state-action frequencies and the expected cumulative reward are rational functions of the policy, whereby the degree is determined by the degree of partial observability. We then describe the optimization problem as a linear optimization problem in the space of feasible state-action frequencies subject to polynomial constraints that we characterize explicitly. This allows us to address the combinatorial and geometric complexity of the optimization problem using recent tools from polynomial optimization. In particular, we demonstrate how the partial observability constraints can lead to multiple smooth and non-smooth local optimizers and we estimate the number of critical points.

(2 min) We propose a machine-translation approach to automatically generate a playlist title from a set of music tracks. We take a sequence of track IDs as input and a sequence of words in a playlist title as output, adapting the sequence-to-sequence framework based on Recurrent Neural Network (RNN) and Transformer to the music data. Considering the orderless nature of music tracks in a playlist, we propose two techniques that remove the order of the input sequence. One is data augmentation by shuffling and the other is deleting the positional encoding. We also reorganize the existing music playlist datasets to generate phrase-level playlist titles. The result shows that the Transformer models generally outperform the RNN model. Also, removing the order of input sequence improves the performance further.

(2 min) Recent advancements in Graph Neural Networks have led to state-of-the-art performance on representation learning of graphs for node classification. However, the majority of existing works process directed graphs by symmetrization, which may cause loss of directional information. In this paper, we propose the magnetic Laplacian that preserves edge directionality by encoding it into complex phase as a deformation of the combinatorial Laplacian. In addition, we design an Auto-Regressive Moving-Average (ARMA) filter that is capable of learning global features from graphs. To reduce time complexity, Taylor expansion is applied to approximate the filter. We derive complex-valued operations in graph neural network and devise a simplified Magnetic Graph Convolution network, namely sMGC. Our experiment results demonstrate that sMGC is a fast, powerful, and widely applicable GNN.

(2 min) In this paper, a marker-based, single-person optical motion capture method (DeepMoCap) is proposed using multiple spatio-temporally aligned infrared-depth sensors and retro-reflective straps and patches (reflectors). DeepMoCap explores motion capture by automatically localizing and labeling reflectors on depth images and, subsequently, on 3D space. Introducing a non-parametric representation to encode the temporal correlation among pairs of colorized depthmaps and 3D optical flow frames, a multi-stage Fully Convolutional Network (FCN) architecture is proposed to jointly learn reflector locations and their temporal dependency among sequential frames. The extracted reflector 2D locations are spatially mapped in 3D space, resulting in robust 3D optical data extraction. The subject's motion is efficiently captured by applying a template-based fitting technique on the extracted optical data. Two datasets have been created and made publicly available for evaluation purposes; one comprising multi-view depth and 3D optical flow annotated images (DMC2.5D), and a second, consisting of spatio-temporally aligned multi-view depth images along with skeleton, inertial and ground truth MoCap data (DMC3D). The FCN model outperforms its competitors on the DMC2.5D dataset using 2D Percentage of Correct Keypoints (PCK) metric, while the motion capture outcome is evaluated against RGB-D and inertial data fusion approaches on DMC3D, outperforming the next best method by 4.5% in total 3D PCK accuracy.

(2 min) We study the problem of estimating a rank-$1$ signal in the presence of rotationally invariant noise-a class of perturbations more general than Gaussian noise. Principal Component Analysis (PCA) provides a natural estimator, and sharp results on its performance have been obtained in the high-dimensional regime. Recently, an Approximate Message Passing (AMP) algorithm has been proposed as an alternative estimator with the potential to improve the accuracy of PCA. However, the existing analysis of AMP requires an initialization that is both correlated with the signal and independent of the noise, which is often unrealistic in practice. In this work, we combine the two methods, and propose to initialize AMP with PCA. Our main result is a rigorous asymptotic characterization of the performance of this estimator. Both the AMP algorithm and its analysis differ from those previously derived in the Gaussian setting: at every iteration, our AMP algorithm requires a specific term to account for PCA initialization, while in the Gaussian case, PCA initialization affects only the first iteration of AMP. The proof is based on a two-phase artificial AMP that first approximates the PCA estimator and then mimics the true AMP. Our numerical simulations show an excellent agreement between AMP results and theoretical predictions, and suggest an interesting open direction on achieving Bayes-optimal performance.

(0 min) Improving adversarial robustness of neural networks remains a major challenge. Fundamentally, training a neural network via gradient descent is a parameter estimation problem. In adaptive control, maintaining persistency of excitation (PoE) is integral to ensuring convergence of parameter estimates in dynamical systems to their true values. We show that parameter estimation with gradient descent can be modeled as a sampling of an adaptive linear time-varying continuous system. Leveraging this model, and with inspiration from Model-Reference Adaptive Control (MRAC), we prove a sufficient condition to constrain gradient descent updates to reference persistently excited trajectories converging to the true parameters. The sufficient condition is achieved when the learning rate is less than the inverse of the Lipschitz constant of the gradient of loss function. We provide an efficient technique for estimating the corresponding Lipschitz constant in practice using extreme value theory. Our experimental results in both standard and adversarial training illustrate that networks trained with the PoE-motivated learning rate schedule have similar clean accuracy but are significantly more robust to adversarial attacks than models trained using current state-of-the-art heuristics.

(0 min) Conventional fine-tuning of pre-trained language models tunes all model parameters and stores a full model copy for each downstream task, which has become increasingly infeasible as the model size grows larger. Recent parameter-efficient language model tuning (PELT) methods manage to match the performance of fine-tuning with much fewer trainable parameters and perform especially well when the training data is limited. However, different PELT methods may perform rather differently on the same task, making it nontrivial to select the most appropriate method for a specific task, especially considering the fast-growing number of new PELT methods and downstream tasks. In light of model diversity and the difficulty of model selection, we propose a unified framework, UniPELT, which incorporates different PELT methods as submodules and learns to activate the ones that best suit the current data or task setup. Remarkably, on the GLUE benchmark, UniPELT consistently achieves 1~3pt gains compared to the best individual PELT method that it incorporates and even outperforms fine-tuning under different setups. Moreover, UniPELT often surpasses the upper bound when taking the best performance of all its submodules used individually on each task, indicating that a mixture of multiple PELT methods may be inherently more effective than single methods.

(0 min) In Byzantine robust distributed or federated learning, a central server wants to train a machine learning model over data distributed across multiple workers. However, a fraction of these workers may deviate from the prescribed algorithm and send arbitrary messages. While this problem has received significant attention recently, most current defenses assume that the workers have identical data. For realistic cases when the data across workers are heterogeneous (non-iid), we design new attacks which circumvent current defenses, leading to significant loss of performance. We then propose a simple bucketing scheme that adapts existing robust algorithms to heterogeneous datasets at a negligible computational cost. We also theoretically and experimentally validate our approach, showing that combining bucketing with existing robust algorithms is effective against challenging attacks. Our work is the first to establish guaranteed convergence for the non-iid Byzantine robust problem under realistic assumptions.

(0 min) Single-cell RNA sequencing (scRNA-seq) has the potential to provide powerful, high-resolution signatures to inform disease prognosis and precision medicine. This paper takes an important first step towards this goal by developing an interpretable machine learning algorithm, CloudPred, to predict individuals' disease phenotypes from their scRNA-seq data. Predicting phenotype from scRNA-seq is challenging for standard machine learning methods -- the number of cells measured can vary by orders of magnitude across individuals and the cell populations are also highly heterogeneous. Typical analysis creates pseudo-bulk samples which are biased toward prior annotations and also lose the single cell resolution. CloudPred addresses these challenges via a novel end-to-end differentiable learning algorithm which is coupled with a biologically informed mixture of cell types model. CloudPred automatically infers the cell subpopulation that are salient for the phenotype without prior annotations. We developed a systematic simulation platform to evaluate the performance of CloudPred and several alternative methods we propose, and find that CloudPred outperforms the alternative methods across several settings. We further validated CloudPred on a real scRNA-seq dataset of 142 lupus patients and controls. CloudPred achieves AUROC of 0.98 while identifying a specific subpopulation of CD4 T cells whose presence is highly indicative of lupus. CloudPred is a powerful new framework to predict clinical phenotypes from scRNA-seq data and to identify relevant cells.

(0 min) Modern software systems and products increasingly rely on machine learning models to make data-driven decisions based on interactions with users and systems, e.g., compute infrastructure. For broader adoption, this practice must (i) accommodate software engineers without ML backgrounds, and (ii) provide mechanisms to optimize for product goals. In this work, we describe general principles and a specific end-to-end ML platform, Looper, which offers easy-to-use APIs for decision-making and feedback collection. Looper supports the full end-to-end ML lifecycle from online data collection to model training, deployment, inference, and extends support to evaluation and tuning against product goals. We outline the platform architecture and overall impact of production deployment. We also describe the learning curve and summarize experiences from platform adopters.

(0 min) Food-choices and eating-habits directly contribute to our long-term health. This makes the food recommender system a potential tool to address the global crisis of obesity and malnutrition. Over the past decade, artificial-intelligence and medical researchers became more invested in researching tools that can guide and help people make healthy and thoughtful decisions around food and diet. In many typical (Recommender System) RS domains, smart nudges have been proven effective in shaping users' consumption patterns. In recent years, knowledgeable nudging and incentifying choices started getting attention in the food domain as well. To develop smart nudging for promoting healthier food choices, we combined Machine Learning and RS technology with food-healthiness guidelines from recognized health organizations, such as the World Health Organization, Food Standards Agency, and the National Health Service United Kingdom. In this paper, we discuss our research on, persuasive visualization for making users aware of the healthiness of the recommended recipes. Here, we propose three novel nudging technology, the WHO-BubbleSlider, the FSA-ColorCoading, and the DRCI-MLCP, that encourage users to choose healthier recipes. We also propose a Topic Modeling based portion-size recommendation algorithm. To evaluate our proposed smart-nudges, we conducted an online user study with 96 participants and 92250 recipes. Results showed that, during the food decision-making process, appropriate healthiness cues make users more likely to click, browse, and choose healthier recipes over less healthy ones.

(0 min) Continuous integration testing is an important step in the modern software engineering life cycle. Test prioritization is a method that can improve the efficiency of continuous integration testing by selecting test cases that can detect faults in the early stage of each cycle. As continuous integration testing produces voluminous test execution data, test history is a commonly used artifact in test prioritization. However, existing test prioritization techniques for continuous integration either cannot handle large test history or are optimized for using a limited number of historical test cycles. We show that such a limitation can decrease fault detection effectiveness of prioritized test suites. This work introduces DeepOrder, a deep learning-based model that works on the basis of regression machine learning. DeepOrder ranks test cases based on the historical record of test executions from any number of previous test cycles. DeepOrder learns failed test cases based on multiple factors including the duration and execution status of test cases. We experimentally show that deep neural networks, as a simple regression model, can be efficiently used for test case prioritization in continuous integration testing. DeepOrder is evaluated with respect to time-effectiveness and fault detection effectiveness in comparison with an industry practice and the state of the art approaches. The results show that DeepOrder outperforms the industry practice and state-of-the-art test prioritization approaches in terms of these two metrics.

(0 min) We study the problem of control policy design for decentralized state-feedback linear quadratic control with a partially nested information structure, when the system model is unknown. We propose a model-based learning solution, which consists of two steps. First, we estimate the unknown system model from a single system trajectory of finite length, using least squares estimation. Next, based on the estimated system model, we design a control policy that satisfies the desired information structure. We show that the suboptimality gap between our control policy and the optimal decentralized control policy (designed using accurate knowledge of the system model) scales linearly with the estimation error of the system model. Using this result, we provide an end-to-end sample complexity result for learning decentralized controllers for a linear quadratic control problem with a partially nested information structure.

(0 min) A challenge in reinforcement learning (RL) is minimizing the cost of sampling associated with exploration. Distributed exploration reduces sampling complexity in multi-agent RL (MARL). We investigate the benefits to performance in MARL when exploration is fully decentralized. Specifically, we consider a class of online, episodic, tabular $Q$-learning problems under time-varying reward and transition dynamics, in which agents can communicate in a decentralized manner.We show that group performance, as measured by the bound on regret, can be significantly improved through communication when each agent uses a decentralized message-passing protocol, even when limited to sending information up to its $\gamma$-hop neighbors. We prove regret and sample complexity bounds that depend on the number of agents, communication network structure and $\gamma.$ We show that incorporating more agents and more information sharing into the group learning scheme speeds up convergence to the optimal policy. Numerical simulations illustrate our results and validate our theoretical claims.

(0 min) Markov chains with variable length are useful parsimonious stochastic models able to generate most stationary sequence of discrete symbols. The idea is to identify the suffixes of the past, called contexts, that are relevant to predict the future symbol. Sometimes a single state is a context, and looking at the past and finding this specific state makes the further past irrelevant. These states are called renewal states and they split the chain into independent blocks. In order to identify renewal states for chains with variable length, we propose the use of Intrinsic Bayes Factor to evaluate the plausibility of each set of renewal states. In this case, the difficulty lies in finding the marginal posterior distribution for the random context trees for general prior distribution on the space of context trees and Dirichlet prior for the transition probabilities. To show the strength of our method, we analyzed artificial datasets generated from two binary models models and one example coming from the field of Linguistics.

(0 min) We introduce a learning-based algorithm to obtain a measurement matrix for compressive sensing related recovery problems. The focus lies on matrices with a constant modulus constraint which typically represent a network of analog phase shifters in hybrid precoding/combining architectures. We interpret a matrix with restricted isometry property as a mapping of points from a high- to a low-dimensional hypersphere. We argue that points on the low-dimensional hypersphere, namely, in the range of the matrix, should be uniformly distributed to increase robustness against measurement noise. This notion is formalized in an optimization problem which uses one of the maximum mean discrepancy metrics in the objective function. Recent success of such metrics in neural network related topics motivate a solution of the problem based on machine learning. Numerical experiments show better performance than random measurement matrices that are generally employed in compressive sensing contexts. Further, we adapt a method from the literature to the constant modulus constraint. This method can also compete with random matrices and it is shown to harmonize well with the proposed learning-based approach if it is used as an initialization. Lastly, we describe how other structural matrix constraints, e.g., a Toeplitz constraint, can be taken into account, too.

(0 min) Selecting powerful predictors for an outcome is a cornerstone task for machine learning. However, some types of questions can only be answered by identifying the predictors that causally affect the outcome. A recent approach to this causal inference problem leverages the invariance property of a causal mechanism across differing experimental environments (Peters et al., 2016; Heinze-Deml et al., 2018). This method, invariant causal prediction (ICP), has a substantial computational defect -- the runtime scales exponentially with the number of possible causal variables. In this work, we show that the approach taken in ICP may be reformulated as a series of nonparametric tests that scales linearly in the number of predictors. Each of these tests relies on the minimization of a novel loss function -- the Wasserstein variance -- that is derived from tools in optimal transport theory and is used to quantify distributional variability across environments. We prove under mild assumptions that our method is able to recover the set of identifiable direct causes, and we demonstrate in our experiments that it is competitive with other benchmark causal discovery algorithms.

(0 min) Many-to-one maps are ubiquitous in machine learning, from the image recognition model that assigns a multitude of distinct images to the concept of "cat" to the time series forecasting model which assigns a range of distinct time-series to a single scalar regression value. While the primary use of such models is naturally to associate correct output to each input, in many problems it is also useful to be able to explore, understand, and sample from a model's fibers, which are the set of input values $x$ such that $f(x) = y$, for fixed $y$ in the output space. In this paper we show that popular generative architectures are ill-suited to such tasks. Motivated by this we introduce a novel generative architecture, a Bundle Network, based on the concept of a fiber bundle from (differential) topology. BundleNets exploit the idea of a local trivialization wherein a space can be locally decomposed into a product space that cleanly encodes the many-to-one nature of the map. By enforcing this decomposition in BundleNets and by utilizing state-of-the-art invertible components, investigating a network's fibers becomes natural.

(0 min) Optimal symbol detection in multiple-input multiple-output (MIMO) systems is known to be an NP-hard problem. Recently, there has been a growing interest to get reasonably close to the optimal solution using neural networks while keeping the computational complexity in check. However, existing work based on deep learning shows that it is difficult to design a generic network that works well for a variety of channels. In this work, we propose a method that tries to strike a balance between symbol error rate (SER) performance and generality of channels. Our method is based on hypernetworks that generate the parameters of a neural network-based detector that works well on a specific channel. We propose a general framework by regularizing the training of the hypernetwork with some pre-trained instances of the channel-specific method. Through numerical experiments, we show that our proposed method yields high performance for a set of prespecified channel realizations while generalizing well to all channels drawn from a specific distribution.

(0 min) The identification of active binding drugs for target proteins (termed as drug-target interaction prediction) is the key challenge in virtual screening, which plays an essential role in drug discovery. Although recent deep learning-based approaches achieved better performance than molecular docking, existing models often neglect certain aspects of the intermolecular information, hindering the performance of prediction. We recognize this problem and propose a novel approach named Intermolecular Graph Transformer (IGT) that employs a dedicated attention mechanism to model intermolecular information with a three-way Transformer-based architecture. IGT outperforms state-of-the-art approaches by 9.1% and 20.5% over the second best for binding activity and binding pose prediction respectively, and shows superior generalization ability to unseen receptor proteins. Furthermore, IGT exhibits promising drug screening ability against SARS-CoV-2 by identifying 83.1% active drugs that have been validated by wet-lab experiments with near-native predicted binding poses.

(0 min) In recent years, deep neural network (DNN) compression systems have proved to be highly effective for designing source codes for many natural sources. However, like many other machine learning systems, these compressors suffer from vulnerabilities to distribution shifts as well as out-of-distribution (OOD) data, which reduces their real-world applications. In this paper, we initiate the study of OOD robust compression. Considering robustness to two types of ambiguity sets (Wasserstein balls and group shifts), we propose algorithmic and architectural frameworks built on two principled methods: one that trains DNN compressors using distributionally-robust optimization (DRO), and the other which uses a structured latent code. Our results demonstrate that both methods enforce robustness compared to a standard DNN compressor, and that using a structured code can be superior to the DRO compressor. We observe tradeoffs between robustness and distortion and corroborate these findings theoretically for a specific class of sources.

(0 min) Recurrent neural networks (RNNs) are wide-spread machine learning tools for modeling sequential and time series data. They are notoriously hard to train because their loss gradients backpropagated in time tend to saturate or diverge during training. This is known as the exploding and vanishing gradient problem. Previous solutions to this issue either built on rather complicated, purpose-engineered architectures with gated memory buffers, or - more recently - imposed constraints that ensure convergence to a fixed point or restrict (the eigenspectrum of) the recurrence matrix. Such constraints, however, convey severe limitations on the expressivity of the RNN. Essential intrinsic dynamics such as multistability or chaos are disabled. This is inherently at disaccord with the chaotic nature of many, if not most, time series encountered in nature and society. Here we offer a comprehensive theoretical treatment of this problem by relating the loss gradients during RNN training to the Lyapunov spectrum of RNN-generated orbits. We mathematically prove that RNNs producing stable equilibrium or cyclic behavior have bounded gradients, whereas the gradients of RNNs with chaotic dynamics always diverge. Based on these analyses and insights, we offer an effective yet simple training technique for chaotic data and guidance on how to choose relevant hyperparameters according to the Lyapunov spectrum.

(0 min) We consider the problem of black-box multi-objective optimization (MOO) using expensive function evaluations (also referred to as experiments), where the goal is to approximate the true Pareto set of solutions by minimizing the total resource cost of experiments. For example, in hardware design optimization, we need to find the designs that trade-off performance, energy, and area overhead using expensive computational simulations. The key challenge is to select the sequence of experiments to uncover high-quality solutions using minimal resources. In this paper, we propose a general framework for solving MOO problems based on the principle of output space entropy (OSE) search: select the experiment that maximizes the information gained per unit resource cost about the true Pareto front. We appropriately instantiate the principle of OSE search to derive efficient algorithms for the following four MOO problem settings: 1) The most basic em single-fidelity setting, where experiments are expensive and accurate; 2) Handling em black-box constraints} which cannot be evaluated without performing experiments; 3) The discrete multi-fidelity setting, where experiments can vary in the amount of resources consumed and their evaluation accuracy; and 4) The em continuous-fidelity setting, where continuous function approximations result in a huge space of experiments. Experiments on diverse synthetic and real-world benchmarks show that our OSE search based algorithms improve over state-of-the-art methods in terms of both computational-efficiency and accuracy of MOO solutions.

(0 min) Recently, the stability of graph filters has been studied as one of the key theoretical properties driving the highly successful graph convolutional neural networks (GCNs). The stability of a graph filter characterizes the effect of topology perturbation on the output of a graph filter, a fundamental building block for GCNs. Many existing results have focused on the regime of small perturbation with a small number of edge rewires. However, the number of edge rewires can be large in many applications. To study the latter case, this work departs from the previous analysis and proves a bound on the stability of graph filter relying on the filter's frequency response. Assuming the graph filter is low pass, we show that the stability of the filter depends on perturbation to the community structure. As an application, we show that for stochastic block model graphs, the graph filter distance converges to zero when the number of nodes approaches infinity. Numerical simulations validate our findings.

(0 min) The Event Horizon Telescope (EHT) recently released the first horizon-scale images of the black hole in M87. Combined with other astronomical data, these images constrain the mass and spin of the hole as well as the accretion rate and magnetic flux trapped on the hole. An important question for the EHT is how well key parameters, such as trapped magnetic flux and the associated disk models, can be extracted from present and future EHT VLBI data products. The process of modeling visibilities and analyzing them is complicated by the fact that the data are sparsely sampled in the Fourier domain while most of the theory/simulation is constructed in the image domain. Here we propose a data-driven approach to analyze complex visibilities and closure quantities for radio interferometric data with neural networks. Using mock interferometric data, we show that our neural networks are able to infer the accretion state as either high magnetic flux (MAD) or low magnetic flux (SANE), suggesting that it is possible to perform parameter extraction directly in the visibility domain without image reconstruction. We have applied VLBInet to real M87 EHT data taken on four different days in 2017 (April 5, 6, 10, 11), and our neural networks give a score prediction 0.52, 0.4, 0.43, 0.76 for each day, with an average score 0.53, which shows no significant indication for the data to lean toward either the MAD or SANE state.

(0 min) Motivated by the success of T5 (Text-To-Text Transfer Transformer) in pre-training natural language processing models, we propose a unified-modal SpeechT5 framework that explores the encoder-decoder pre-training for self-supervised speech/text representation learning. The SpeechT5 framework consists of a shared encoder-decoder network and six modal-specific (speech/text) pre/post-nets. After preprocessing the speech/text input through the pre-nets, the shared encoder-decoder network models the sequence to sequence transformation, and then the post-nets generate the output in the speech/text modality based on the decoder output. Particularly, SpeechT5 can pre-train on a large scale of unlabeled speech and text data to improve the capability of the speech and textual modeling. To align the textual and speech information into a unified semantic space, we propose a cross-modal vector quantization method with random mixing-up to bridge speech and text. Extensive evaluations on a wide variety of spoken language processing tasks, including voice conversion, automatic speech recognition, text to speech, and speaker identification, show the superiority of the proposed SpeechT5 framework.

(0 min) Development of new machine learning models is typically done on manually curated data sets, making them unsuitable for evaluating the models' performance during operations, where the evaluation needs to be performed automatically on incoming streams of new data. Unfortunately, pure reliance on a fully automatic pipeline for monitoring model performance makes it difficult to understand if any observed performance issues are due to model performance, pipeline issues, emerging data distribution biases, or some combination of the above. With this in mind, we developed a web-based visualization system that allows the users to quickly gather headline performance numbers while maintaining confidence that the underlying data pipeline is functioning properly. It also enables the users to immediately observe the root cause of an issue when something goes wrong. We introduce a novel way to analyze performance under data issues using a data coverage equalizer. We describe the various modifications and additional plots, filters, and drill-downs that we added on top of the standard evaluation metrics typically tracked in machine learning (ML) applications, and walk through some real world examples that proved valuable for introspecting our models.

(0 min) Text autoencoders are often used for unsupervised conditional text generation by applying mappings in the latent space to change attributes to the desired values. Recently, Mai et al. (2020) proposed Emb2Emb, a method to learn these mappings in the embedding space of an autoencoder. However, their method is restricted to autoencoders with a single-vector embedding, which limits how much information can be retained. We address this issue by extending their method to Bag-of-Vectors Autoencoders (BoV-AEs), which encode the text into a variable-size bag of vectors that grows with the size of the text, as in attention-based models. This allows to encode and reconstruct much longer texts than standard autoencoders. Analogous to conventional autoencoders, we propose regularization techniques that facilitate learning meaningful operations in the latent space. Finally, we adapt for a training scheme that learns to map an input bag to an output bag, including a novel loss function and neural architecture. Our experimental evaluations on unsupervised sentiment transfer and sentence summarization show that our method performs substantially better than a standard autoencoder.

(0 min) The gradient descent-ascent (GDA) algorithm has been widely applied to solve nonconvex minimax optimization problems. However, the existing GDA-type algorithms can only find first-order stationary points of the envelope function of nonconvex minimax optimization problems, which does not rule out the possibility to get stuck at suboptimal saddle points. In this paper, we develop Cubic-GDA -- the first GDA-type algorithm for escaping strict saddle points in nonconvex-strongly-concave minimax optimization. Specifically, the algorithm uses gradient ascent to estimate the second-order information of the minimax objective function, and it leverages the cubic regularization technique to efficiently escape the strict saddle points. Under standard smoothness assumptions on the objective function, we show that Cubic-GDA admits an intrinsic potential function whose value monotonically decreases in the minimax optimization process. Such a property leads to a desired global convergence of Cubic-GDA to a second-order stationary point at a sublinear rate. Moreover, we analyze the convergence rate of Cubic-GDA in the full spectrum of a gradient dominant-type nonconvex geometry. Our result shows that Cubic-GDA achieves an orderwise faster convergence rate than the standard GDA for a wide spectrum of gradient dominant geometry. Our study bridges minimax optimization with second-order optimization and may inspire new developments along this direction.

(0 min) In medical dialogue summarization, summaries must be coherent and must capture all the medically relevant information in the dialogue. However, learning effective models for summarization require large amounts of labeled data which is especially hard to obtain. We present an algorithm to create synthetic training data with an explicit focus on capturing medically relevant information. We utilize GPT-3 as the backbone of our algorithm and scale 210 human labeled examples to yield results comparable to using 6400 human labeled examples (~30x) leveraging low-shot learning and an ensemble method. In detailed experiments, we show that this approach produces high quality training data that can further be combined with human labeled data to get summaries that are strongly preferable to those produced by models trained on human data alone both in terms of medical accuracy and coherency.

(0 min) Neural topic models (NTMs) apply deep neural networks to topic modelling. Despite their success, NTMs generally ignore two important aspects: (1) only document-level word count information is utilized for the training, while more fine-grained sentence-level information is ignored, and (2) external semantic knowledge regarding documents, sentences and words are not exploited for the training. To address these issues, we propose a variational autoencoder (VAE) NTM model that jointly reconstructs the sentence and document word counts using combinations of bag-of-words (BoW) topical embeddings and pre-trained semantic embeddings. The pre-trained embeddings are first transformed into a common latent topical space to align their semantics with the BoW embeddings. Our model also features hierarchical KL divergence to leverage embeddings of each document to regularize those of their sentences, thereby paying more attention to semantically relevant sentences. Both quantitative and qualitative experiments have shown the efficacy of our model in 1) lowering the reconstruction errors at both the sentence and document levels, and 2) discovering more coherent topics from real-world datasets.

(0 min) Standard models of biologically realistic, or inspired, reinforcement learning employ a global error signal which implies shallow networks. However, deep networks could offer a drastically superior performance by feeding the error signal backwards through such a network which in turn is not biologically realistic as it requires symmetric weights between top-down and bottom-up pathways. Instead, we present a network combining local learning with global modulation where neuromodulation controls the amount of plasticity change in the whole network, while only the sign of the error is backpropagated through the network. The neuromodulation can be understood as a rectified error, or relevance, signal while the bottom-up sign of the error signal decides between long-term potentiation and long-term depression. We demonstrate the performance of this paradigm with a real robotic task.

(0 min) Recently, physics informed neural networks have successfully been applied to a broad variety of problems in applied mathematics and engineering. The principle idea is to use a neural network as a global ansatz function to partial differential equations. Due to the global approximation, physics informed neural networks have difficulties in displaying localized effects and strong non-linear solutions by optimization. In this work we consider material non-linearities invoked by material inhomogeneities with sharp phase interfaces. This constitutes a challenging problem for a method relying on a global ansatz. To overcome convergence issues, adaptive training strategies and domain decomposition are studied. It is shown, that the domain decomposition approach is able to accurately resolve nonlinear stress, displacement and energy fields in heterogeneous microstructures obtained from real-world $\mu$CT-scans.

(0 min) In collaborative or federated learning, model personalization can be a very effective strategy to deal with heterogeneous training data across clients. We introduce WAFFLE (Weighted Averaging For Federated LEarning), a personalized collaborative machine learning algorithm based on SCAFFOLD. SCAFFOLD uses stochastic control variates to converge towards a model close to the globally optimal model even in tasks where the distribution of data and labels across clients is highly skewed. In contrast, WAFFLE uses the Euclidean distance between clients' updates to weigh their individual contributions and thus minimize the trained personalized model loss on the specific agent of interest. Through a series of experiments, we compare our proposed new method to two recent personalized federated learning methods, Weight Erosion and APFL, as well as two global learning methods, federated averaging and SCAFFOLD. We evaluate our method using two categories of non-identical client data distributions (concept shift and label skew) on two benchmark image data sets, MNIST and CIFAR10. Our experiments demonstrate the effectiveness of WAFFLE compared with other methods, as it achieves or improves accuracy with faster convergence.

(0 min) Many methods to explain black-box models, whether local or global, are additive. In this paper, we study global additive explanations for non-additive models, focusing on four explanation methods: partial dependence, Shapley explanations adapted to a global setting, distilled additive explanations, and gradient-based explanations. We show that different explanation methods characterize non-additive components in a black-box model's prediction function in different ways. We use the concepts of main and total effects to anchor additive explanations, and quantitatively evaluate additive and non-additive explanations. Even though distilled explanations are generally the most accurate additive explanations, non-additive explanations such as tree explanations that explicitly model non-additive components tend to be even more accurate. Despite this, our user study showed that machine learning practitioners were better able to leverage additive explanations for various tasks. These considerations should be taken into account when considering which explanation to trust and use to explain black-box models.

(0 min) Transformers have revolutionized the world of deep learning, specially in the field of natural language processing. Recently, the Audio Spectrogram Transformer (AST) was proposed for audio classification, leading to state of the art results in several datasets. However, in order for ASTs to outperform CNNs, pretraining with ImageNet is needed. In this paper, we study one component of the AST, the positional encoding, and propose several variants to improve the performance of ASTs trained from scratch, without ImageNet pretraining. Our best model, which incorporates conditional positional encodings, significantly improves performance on Audioset and ESC-50 compared to the original AST.

(0 min) With the growth of the academic engines, the mining and analysis acquisition of massive researcher data, such as collaborator recommendation and researcher retrieval, has become indispensable. It can improve the quality of services and intelligence of academic engines. Most of the existing studies for researcher data mining focus on a single task for a particular application scenario and learning a task-specific model, which is usually unable to transfer to out-of-scope tasks. The pre-training technology provides a generalized and sharing model to capture valuable information from enormous unlabeled data. The model can accomplish multiple downstream tasks via a few fine-tuning steps. In this paper, we propose a multi-task self-supervised learning-based researcher data pre-training model named RPT. Specifically, we divide the researchers' data into semantic document sets and community graph. We design the hierarchical Transformer and the local community encoder to capture information from the two categories of data, respectively. Then, we propose three self-supervised learning objectives to train the whole model. Finally, we also propose two transfer modes of RPT for fine-tuning in different scenarios. We conduct extensive experiments to evaluate RPT, results on three downstream tasks verify the effectiveness of pre-training for researcher data mining.

(0 min) We study a new family of inverse problems for recovering representations of corrupted data. We assume access to a pre-trained representation learning network R(x) that operates on clean images, like CLIP. The problem is to recover the representation of an image R(x), if we are only given a corrupted version A(x), for some known forward operator A. We propose a supervised inversion method that uses a contrastive objective to obtain excellent representations for highly corrupted images. Using a linear probe on our robust representations, we achieve a higher accuracy than end-to-end supervised baselines when classifying images with various types of distortions, including blurring, additive noise, and random pixel masking. We evaluate on a subset of ImageNet and observe that our method is robust to varying levels of distortion. Our method outperforms end-to-end baselines even with a fraction of the labeled data in a wide range of forward operators.

(0 min) In black-box optimization problems, we aim to maximize an unknown objective function, where the function is only accessible through feedbacks of an evaluation or simulation oracle. In real-life, the feedbacks of such oracles are often noisy and available after some unknown delay that may depend on the computation time of the oracle. Additionally, if the exact evaluations are expensive but coarse approximations are available at a lower cost, the feedbacks can have multi-fidelity. In order to address this problem, we propose a generic extension of hierarchical optimistic tree search (HOO), called ProCrastinated Tree Search (PCTS), that flexibly accommodates a delay and noise-tolerant bandit algorithm. We provide a generic proof technique to quantify regret of PCTS under delayed, noisy, and multi-fidelity feedbacks. Specifically, we derive regret bounds of PCTS enabled with delayed-UCB1 (DUCB1) and delayed-UCB-V (DUCBV) algorithms. Given a horizon $T$, PCTS retains the regret bound of non-delayed HOO for expected delay of $O(\log T)$ and worsens by $O(T^{\frac{1-\alpha}{d+2}})$ for expected delays of $O(T^{1-\alpha})$ for $\alpha \in (0,1]$. We experimentally validate on multiple synthetic functions and hyperparameter tuning problems that PCTS outperforms the state-of-the-art black-box optimization methods for feedbacks with different noise levels, delays, and fidelity.

(0 min) Kinship verification is the task of determining whether a parent-child, sibling, or grandparent-grandchild relationship exists between two people and is important in social media applications, forensic investigations, finding missing children, and reuniting families. We demonstrate high quality kinship verification by participating in the FG 2021 Recognizing Families in the Wild challenge which provides the largest publicly available dataset in the field. Our approach is among the top 3 winning entries in the competition. We ensemble models written by both human experts and OpenAI Codex. We make our models and code publicly available.

(0 min) In analyzing complex datasets, it is often of interest to infer lower dimensional structure underlying the higher dimensional observations. As a flexible class of nonlinear structures, it is common to focus on Riemannian manifolds. Most existing manifold learning algorithms replace the original data with lower dimensional coordinates without providing an estimate of the manifold in the observation space or using the manifold to denoise the original data. This article proposes a new methodology for addressing these problems, allowing interpolation of the estimated manifold between fitted data points. The proposed approach is motivated by novel theoretical properties of local covariance matrices constructed from noisy samples on a manifold. Our results enable us to turn a global manifold reconstruction problem into a local regression problem, allowing application of Gaussian processes for probabilistic manifold reconstruction. In addition to theory justifying the algorithm, we provide simulated and real data examples to illustrate the performance.

(0 min) Graph embedding techniques have attracted growing interest since they convert the graph data into continuous and low-dimensional space. Effective graph analytic provides users a deeper understanding of what is behind the data and thus can benefit a variety of machine learning tasks. With the current scale of real-world applications, most graph analytic methods suffer high computation and space costs. These methods and systems can process a network with thousands to a few million nodes. However, scaling to large-scale networks remains a challenge. The complexity of training graph embedding system requires the use of existing accelerators such as GPU. In this paper, we introduce a hybrid CPU-GPU framework that addresses the challenges of learning embedding of large-scale graphs. The performance of our method is compared qualitatively and quantitatively with the existing embedding systems on common benchmarks. We also show that our system can scale training to datasets with an order of magnitude greater than a single machine's total memory capacity. The effectiveness of the learned embedding is evaluated within multiple downstream applications. The experimental results indicate the effectiveness of the learned embedding in terms of performance and accuracy.

(0 min) This study aims to explore different pre-trained models offered in the Torchvision package which is available in the PyTorch library. And investigate their effectiveness on fine-grained images classification. Transfer Learning is an effective method of achieving extremely good performance with insufficient training data. In many real-world situations, people cannot collect sufficient data required to train a deep neural network model efficiently. Transfer Learning models are pre-trained on a large data set, and can bring a good performance on smaller datasets with significantly lower training time. Torchvision package offers us many models to apply the Transfer Learning on smaller datasets. Therefore, researchers may need a guideline for the selection of a good model. We investigate Torchvision pre-trained models on four different data sets: 10 Monkey Species, 225 Bird Species, Fruits 360, and Oxford 102 Flowers. These data sets have images of different resolutions, class numbers, and different achievable accuracies. We also apply their usual fully-connected layer and the Spinal fully-connected layer to investigate the effectiveness of SpinalNet. The Spinal fully-connected layer brings better performance in most situations. We apply the same augmentation for different models for the same data set for a fair comparison. This paper may help future Computer Vision researchers in choosing a proper Transfer Learning model.

(0 min) Zero-shot learning (ZSL) aims to recognize unseen classes based on the knowledge of seen classes. Previous methods focused on learning direct embeddings from global features to the semantic space in hope of knowledge transfer from seen classes to unseen classes. However, an unseen class shares local visual features with a set of seen classes and leveraging global visual features makes the knowledge transfer ineffective. To tackle this problem, we propose a Region Semantically Aligned Network (RSAN), which maps local features of unseen classes to their semantic attributes. Instead of using global features which are obtained by an average pooling layer after an image encoder, we directly utilize the output of the image encoder which maintains local information of the image. Concretely, we obtain each attribute from a specific region of the output and exploit these attributes for recognition. As a result, the knowledge of seen classes can be successfully transferred to unseen classes in a region-bases manner. In addition, we regularize the image encoder through attribute regression with a semantic knowledge to extract robust and attribute-related visual features. Experiments on several standard ZSL datasets reveal the benefit of the proposed RSAN method, outperforming state-of-the-art methods.

(0 min) Recent methods for embodied instruction following are typically trained end-to-end using imitation learning. This requires the use of expert trajectories and low-level language instructions. Such approaches assume learned hidden states will simultaneously integrate semantics from the language and vision to perform state tracking, spatial memory, exploration, and long-term planning. In contrast, we propose a modular method with structured representations that (1) builds a semantic map of the scene, and (2) performs exploration with a semantic search policy, to achieve the natural language goal. Our modular method achieves SOTA performance (24.46%) with a substantial (8.17 % absolute) gap from previous work while using less data by eschewing both expert trajectories and low-level instructions. Leveraging low-level language, however, can further increase our performance (26.49%). Our findings suggest that an explicit spatial memory and a semantic search policy can provide a stronger and more general representation for state-tracking and guidance, even in the absence of expert trajectories or low-level instructions.

(0 min) Convolutional neural networks applied for real-world classification tasks need to recognize inputs that are far or out-of-distribution (OoD) with respect to the known or training data. To achieve this, many methods estimate class-conditional posterior probabilities and use confidence scores obtained from the posterior distributions. Recent works propose to use multivariate Gaussian distributions as models of posterior distributions at different layers of the CNN (i.e., for low- and upper-level features), which leads to the confidence scores based on the Mahalanobis distance. However, this procedure involves estimating probability density in high dimensional data using the insufficient number of observations (e.g. the dimensionality of features at the last two layers in the ResNet-101 model are 2048 and 1024, with ca. 1000 observations per class used to estimate density). In this work, we want to address this problem. We show that in many OoD studies in high-dimensional data, LOF-based (Local Outlierness-Factor) methods outperform the parametric, Mahalanobis distance-based methods. This motivates us to propose the nonparametric, LOF-based method of generating the confidence scores for CNNs. We performed several feasibility studies involving ResNet-101 and EffcientNet-B3, based on CIFAR-10 and ImageNet (as known data), and CIFAR-100, SVHN, ImageNet2010, Places365, or ImageNet-O (as outliers). We demonstrated that nonparametric LOF-based confidence estimation can improve current Mahalanobis-based SOTA or obtain similar performance in a simpler way.

(0 min) Graph neural networks (GNNs) and label propagation represent two interrelated modeling strategies designed to exploit graph structure in tasks such as node property prediction. The former is typically based on stacked message-passing layers that share neighborhood information to transform node features into predictive embeddings. In contrast, the latter involves spreading label information to unlabeled nodes via a parameter-free diffusion process, but operates independently of the node features. Given then that the material difference is merely whether features or labels are smoothed across the graph, it is natural to consider combinations of the two for improving performance. In this regard, it has recently been proposed to use a randomly-selected portion of the training labels as GNN inputs, concatenated with the original node features for making predictions on the remaining labels. This so-called label trick accommodates the parallel use of features and labels, and is foundational to many of the top-ranking submissions on the Open Graph Benchmark (OGB) leaderboard. And yet despite its wide-spread adoption, thus far there has been little attempt to carefully unpack exactly what statistical properties the label trick introduces into the training pipeline, intended or otherwise. To this end, we prove that under certain simplifying assumptions, the stochastic label trick can be reduced to an interpretable, deterministic training objective composed of two factors. The first is a data-fitting term that naturally resolves potential label leakage issues, while the second serves as a regularization factor conditioned on graph structure that adapts to graph size and connectivity. Later, we leverage this perspective to motivate a broader range of label trick use cases, and provide experiments to verify the efficacy of these extensions.

(0 min) In this paper, we propose a new deep unfolding neural network based on the ADMM algorithm for analysis Compressed Sensing. The proposed network jointly learns a redundant analysis operator for sparsification and reconstructs the signal of interest. We compare our proposed network with a state-of-the-art unfolded ISTA decoder, that also learns an orthogonal sparsifier. Moreover, we consider not only image, but also speech datasets as test examples. Computational experiments demonstrate that our proposed network outperforms the state-of-the-art deep unfolding networks, consistently for both real-world image and speech datasets.

(0 min) Industrial Safety deals with the physical integrity of humans, machines and the environment when they interact during production scenarios. Industrial Safety is subject to a rigorous certification process that leads to inflexible settings, in which all changes are forbidden. With the progressing introduction of smart robotics and smart machinery to the factory floor, combined with an increasing shortage of skilled workers, it becomes imperative that safety scenarios incorporate a flexible handling of the boundary between humans, machines and the environment. In order to increase the well-being of workers, reduce accidents, and compensate for different skill sets, the configuration of machines and the factory floor should be dynamically adapted, while still enforcing functional safety requirements. The contribution of this paper is as follows: (1) We present a set of three scenarios, and discuss how industrial safety mechanisms could be augmented through dynamic changes to the work environment in order to decrease potential accidents, and thus increase productivity. (2) We introduce the concept of a Cognition Aware Safety System (CASS) and its architecture. The idea behind CASS is to integrate AI based reasoning about human load, stress, and attention with AI based selection of actions to avoid the triggering of safety stops. (3) And finally, we will describe the required performance measurement dimensions for a quantitative performance measurement model to enable a comprehensive (triple bottom line) impact assessment of CASS. Additionally we introduce a detailed guideline for expert interviews to explore the feasibility of the approach for given scenarios.

(0 min) Optimal transport is a framework for comparing measures whereby a cost is incurred for transporting one measure to another. Recent works have aimed to improve optimal transport plans through the introduction of various forms of structure. We introduce novel order constraints into the optimal transport formulation to allow for the incorporation of structure. While there will are now quadratically many constraints as before, we prove a $\delta-$approximate solution to the order-constrained optimal transport problem can be obtained in $\mathcal{O}(L^2\delta^{-2} \kappa(\delta(2cL_\infty (1+(mn)^{1/2}))^{-1}) \cdot mn\log mn)$ time. We derive computationally efficient lower bounds that allow for an explainable approach to adding structure to the optimal transport plan through order constraints. We demonstrate experimentally that order constraints improve explainability using the e-SNLI (Stanford Natural Language Inference) dataset that includes human-annotated rationales for each assignment.

(0 min) Few-shot class incremental learning -- the problem of updating a trained classifier to discriminate among an expanded set of classes with limited labeled data -- is a key challenge for machine learning systems deployed in non-stationary environments. Existing approaches to the problem rely on complex model architectures and training procedures that are difficult to tune and re-use. In this paper, we present an extremely simple approach that enables the use of ordinary logistic regression classifiers for few-shot incremental learning. The key to this approach is a new family of subspace regularization schemes that encourage weight vectors for new classes to lie close to the subspace spanned by the weights of existing classes. When combined with pretrained convolutional feature extractors, logistic regression models trained with subspace regularization outperform specialized, state-of-the-art approaches to few-shot incremental image classification by up to 22% on the miniImageNet dataset. Because of its simplicity, subspace regularization can be straightforwardly extended to incorporate additional background information about the new classes (including class names and descriptions specified in natural language); these further improve accuracy by up to 2%. Our results show that simple geometric regularization of class representations offers an effective tool for continual learning.

(0 min) The nature of the Fermi gamma-ray Galactic Center Excess (GCE) has remained a persistent mystery for over a decade. Although the excess is broadly compatible with emission expected due to dark matter annihilation, an explanation in terms of a population of unresolved astrophysical point sources e.g., millisecond pulsars, remains viable. The effort to uncover the origin of the GCE is hampered in particular by an incomplete understanding of diffuse emission of Galactic origin. This can lead to spurious features that make it difficult to robustly differentiate smooth emission, as expected for a dark matter origin, from more "clumpy" emission expected for a population of relatively bright, unresolved point sources. We use recent advancements in the field of simulation-based inference, in particular density estimation techniques using normalizing flows, in order to characterize the contribution of modeled components, including unresolved point source populations, to the GCE. Compared to traditional techniques based on the statistical distribution of photon counts, our machine learning-based method is able to utilize more of the information contained in a given model of the Galactic Center emission, and in particular can perform posterior parameter estimation while accounting for pixel-to-pixel spatial correlations in the gamma-ray map. This makes the method demonstrably more resilient to certain forms of model misspecification. On application to Fermi data, the method generically attributes a smaller fraction of the GCE flux to unresolved point sources when compared to traditional approaches. We nevertheless infer such a contribution to make up a non-negligible fraction of the GCE across all analysis variations considered, with at least $38^{+9}_{-19}\%$ of the excess attributed to unresolved points sources in our baseline analysis.

(0 min) Federated learning (FL) is a promising distributed learning technique particularly suitable for wireless learning scenarios since it can accomplish a learning task without raw data transportation so as to preserve data privacy and lower network resource consumption. However, current works on FL over wireless communication do not profoundly study the fundamental performance of FL that suffers from data delivery outage due to network interference and data heterogeneity among mobile clients. To accurately exploit the performance of FL over wireless communication, this paper proposes a new FL model over a cellular-connected unmanned aerial vehicle (UAV) network, which characterizes data delivery outage from UAV clients to their server and data heterogeneity among the datasets of UAV clients. We devise a simulation-based approach to evaluating the convergence performance of the proposed FL model. We then propose a tractable analytical framework of the uplink outage probability in the cellular-connected UAV network and derive a neat expression of the uplink outage probability, which reveals how the proposed FL model is impacted by data delivery outage and UAV deployment. Extensive numerical simulations are conducted to show the consistency between the estimated and simulated performances.

(0 min) Continual learning (CL) aims to learn a sequence of tasks without forgetting the previously acquired knowledge. However, recent advances in continual learning are restricted to supervised continual learning (SCL) scenarios. Consequently, they are not scalable to real-world applications where the data distribution is often biased and unannotated. In this work, we focus on unsupervised continual learning (UCL), where we learn the feature representations on an unlabelled sequence of tasks and show that reliance on annotated data is not necessary for continual learning. We conduct a systematic study analyzing the learned feature representations and show that unsupervised visual representations are surprisingly more robust to catastrophic forgetting, consistently achieve better performance, and generalize better to out-of-distribution tasks than SCL. Furthermore, we find that UCL achieves a smoother loss landscape through qualitative analysis of the learned representations and learns meaningful feature representations. Additionally, we propose Lifelong Unsupervised Mixup (LUMP), a simple yet effective technique that leverages the interpolation between the current task and previous tasks' instances to alleviate catastrophic forgetting for unsupervised representations.

(0 min) We introduce HUMAN4D, a large and multimodal 4D dataset that contains a variety of human activities simultaneously captured by a professional marker-based MoCap, a volumetric capture and an audio recording system. By capturing 2 female and $2$ male professional actors performing various full-body movements and expressions, HUMAN4D provides a diverse set of motions and poses encountered as part of single- and multi-person daily, physical and social activities (jumping, dancing, etc.), along with multi-RGBD (mRGBD), volumetric and audio data. Despite the existence of multi-view color datasets captured with the use of hardware (HW) synchronization, to the best of our knowledge, HUMAN4D is the first and only public resource that provides volumetric depth maps with high synchronization precision due to the use of intra- and inter-sensor HW-SYNC. Moreover, a spatio-temporally aligned scanned and rigged 3D character complements HUMAN4D to enable joint research on time-varying and high-quality dynamic meshes. We provide evaluation baselines by benchmarking HUMAN4D with state-of-the-art human pose estimation and 3D compression methods. For the former, we apply 2D and 3D pose estimation algorithms both on single- and multi-view data cues. For the latter, we benchmark open-source 3D codecs on volumetric data respecting online volumetric video encoding and steady bit-rates. Furthermore, qualitative and quantitative visual comparison between mesh-based volumetric data reconstructed in different qualities showcases the available options with respect to 4D representations. HUMAN4D is introduced to the computer vision and graphics research communities to enable joint research on spatio-temporally aligned pose, volumetric, mRGBD and audio data cues. The dataset and its code are available https://tofis.github.io/myurls/human4d.

(0 min) Statistical shape modeling (SSM) characterizes anatomical variations in a population of shapes generated from medical images. SSM requires consistent shape representation across samples in shape cohort. Establishing this representation entails a processing pipeline that includes anatomy segmentation, re-sampling, registration, and non-linear optimization. These shape representations are then used to extract low-dimensional shape descriptors that facilitate subsequent analyses in different applications. However, the current process of obtaining these shape descriptors from imaging data relies on human and computational resources, requiring domain expertise for segmenting anatomies of interest. Moreover, this same taxing pipeline needs to be repeated to infer shape descriptors for new image data using a pre-trained/existing shape model. Here, we propose DeepSSM, a deep learning-based framework for learning the functional mapping from images to low-dimensional shape descriptors and their associated shape representations, thereby inferring statistical representation of anatomy directly from 3D images. Once trained using an existing shape model, DeepSSM circumvents the heavy and manual pre-processing and segmentation and significantly improves the computational time, making it a viable solution for fully end-to-end SSM applications. In addition, we introduce a model-based data-augmentation strategy to address data scarcity. Finally, this paper presents and analyzes two different architectural variants of DeepSSM with different loss functions using three medical datasets and their downstream clinical application. Experiments showcase that DeepSSM performs comparably or better to the state-of-the-art SSM both quantitatively and on application-driven downstream tasks. Therefore, DeepSSM aims to provide a comprehensive blueprint for deep learning-based image-to-shape models.

(0 min) In reinforcement learning (RL), when defining a Markov Decision Process (MDP), the environment dynamics is implicitly assumed to be stationary. This assumption of stationarity, while simplifying, can be unrealistic in many scenarios. In the continual reinforcement learning scenario, the sequence of tasks is another source of nonstationarity. In this work, we propose to examine this continual reinforcement learning setting through the block contextual MDP (BC-MDP) framework, which enables us to relax the assumption of stationarity. This framework challenges RL algorithms to handle both nonstationarity and rich observation settings and, by additionally leveraging smoothness properties, enables us to study generalization bounds for this setting. Finally, we take inspiration from adaptive control to propose a novel algorithm that addresses the challenges introduced by this more realistic BC-MDP setting, allows for zero-shot adaptation at evaluation time, and achieves strong performance on several nonstationary environments.

(0 min) We introduce a new type of format-transforming encryption where the format of ciphertexts is implicitly encoded within a machine-learned generative model. Around this primitive, we build a system for covert messaging over large, public internet platforms (e.g., Twitter). Loosely, our system composes an authenticated encryption scheme, with a method for encoding random ciphertext bits into samples from the generative model's family of seed-indexed token-distributions. By fixing a deployment scenario, we are forced to consider system-level and algorithmic solutions to real challenges -- such as receiver-side parsing ambiguities, and the low information-carrying capacity of actual token-distributions -- that were elided in prior work. We use GPT-2 as our generative model so that our system cryptographically transforms plaintext bitstrings into natural-language covertexts suitable for posting to public platforms. We consider adversaries with full view of the internet platform's content, whose goal is to surface posts that are using our system for covert messaging. We carry out a suite of experiments to provide heuristic evidence of security and to explore tradeoffs between operational efficiency and detectability.

(0 min) We propose a learning-based robust predictive control algorithm that compensates for significant uncertainty in the dynamics for a class of discrete-time systems that are nominally linear with an additive nonlinear component. Such systems commonly model the nonlinear effects of an unknown environment on a nominal system. We optimize over a class of nonlinear feedback policies inspired by certainty equivalent "estimate-and-cancel" control laws pioneered in classical adaptive control to achieve significant performance improvements in the presence of uncertainties of large magnitude, a setting in which existing learning-based predictive control algorithms often struggle to guarantee safety. In contrast to previous work in robust adaptive MPC, our approach allows us to take advantage of structure (i.e., the numerical predictions) in the a priori unknown dynamics learned online through function approximation. Our approach also extends typical nonlinear adaptive control methods to systems with state and input constraints even when we cannot directly cancel the additive uncertain function from the dynamics. Moreover, we apply contemporary statistical estimation techniques to certify the system's safety through persistent constraint satisfaction with high probability. Finally, we show in simulation that our method can accommodate more significant unknown dynamics terms than existing methods.

(0 min) This paper presents an extension and an elaboration of the theory of differential similarity, which was originally proposed in arXiv:1401.2411 [cs.LG]. The goal is to develop an algorithm for clustering and coding that combines a geometric model with a probabilistic model in a principled way. For simplicity, the geometric model in the earlier paper was restricted to the three-dimensional case. The present paper removes this restriction, and considers the full $n$-dimensional case. Although the mathematical model is the same, the strategies for computing solutions in the $n$-dimensional case are different, and one of the main purposes of this paper is to develop and analyze these strategies. Another main purpose is to devise techniques for estimating the parameters of the model from sample data, again in $n$ dimensions. We evaluate the solution strategies and the estimation techniques by applying them to two familiar real-world examples: the classical MNIST dataset and the CIFAR-10 dataset.

(0 min) We develop new parameter-free and scale-free algorithms for solving convex-concave saddle-point problems. Our results are based on a new simple regret minimizer, the Conic Blackwell Algorithm$^+$ (CBA$^+$), which attains $O(1/\sqrt{T})$ average regret. Intuitively, our approach generalizes to other decision sets of interest ideas from the Counterfactual Regret minimization (CFR$^+$) algorithm, which has very strong practical performance for solving sequential games on simplexes. We show how to implement CBA$^+$ for the simplex, $\ell_{p}$ norm balls, and ellipsoidal confidence regions in the simplex, and we present numerical experiments for solving matrix games and distributionally robust optimization problems. Our empirical results show that CBA$^+$ is a simple algorithm that outperforms state-of-the-art methods on synthetic data and real data instances, without the need for any choice of step sizes or other algorithmic parameters.

(0 min) Empirical science of neural scaling laws is a rapidly growing area of significant importance to the future of machine learning, particularly in the light of recent breakthroughs achieved by large-scale pre-trained models such as GPT-3, CLIP and DALL-e. Accurately predicting the neural network performance with increasing resources such as data, compute and model size provides a more comprehensive evaluation of different approaches across multiple scales, as opposed to traditional point-wise comparisons of fixed-size models on fixed-size benchmarks, and, most importantly, allows for focus on the best-scaling, and thus most promising in the future, approaches. In this work, we consider a challenging problem of few-shot learning in image classification, especially when the target data distribution in the few-shot phase is different from the source, training, data distribution, in a sense that it includes new image classes not encountered during training. Our current main goal is to investigate how the amount of pre-training data affects the few-shot generalization performance of standard image classifiers. Our key observations are that (1) such performance improvements are well-approximated by power laws (linear log-log plots) as the training set size increases, (2) this applies to both cases of target data coming from either the same or from a different domain (i.e., new classes) as the training data, and (3) few-shot performance on new classes converges at a faster rate than the standard classification performance on previously seen classes. Our findings shed new light on the relationship between scale and generalization.

(0 min) The matrix normal model, the family of Gaussian matrix-variate distributions whose covariance matrix is the Kronecker product of two lower dimensional factors, is frequently used to model matrix-variate data. The tensor normal model generalizes this family to Kronecker products of three or more factors. We study the estimation of the Kronecker factors of the covariance matrix in the matrix and tensor models. We show nonasymptotic bounds for the error achieved by the maximum likelihood estimator (MLE) in several natural metrics. In contrast to existing bounds, our results do not rely on the factors being well-conditioned or sparse. For the matrix normal model, all our bounds are minimax optimal up to logarithmic factors, and for the tensor normal model our bound for the largest factor and overall covariance matrix are minimax optimal up to constant factors provided there are enough samples for any estimator to obtain constant Frobenius error. In the same regimes as our sample complexity bounds, we show that an iterative procedure to compute the MLE known as the flip-flop algorithm converges linearly with high probability. Our main tool is geodesic strong convexity in the geometry on positive-definite matrices induced by the Fisher information metric. This strong convexity is determined by the expansion of certain random quantum channels. We also provide numerical evidence that combining the flip-flop algorithm with a simple shrinkage estimator can improve performance in the undersampled regime.

(0 min) Neurophysiological time series, such as electromyographic signal and intracortical recordings, are typically composed of many individual spiking sources, the recovery of which can give fundamental insights into the biological system of interest or provide neural information for man-machine interfaces. For this reason, source separation algorithms have become an increasingly important tool in neuroscience and neuroengineering. However, in noisy or highly multivariate recordings these decomposition techniques often make a large number of errors, which degrades human-machine interfacing applications and often requires costly post-hoc manual cleaning of the output label set of spike timestamps. To address both the need for automated post-hoc cleaning and robust separation filters we propose a methodology based on deep metric learning, using a novel loss function which maintains intra-class variance, creating a rich embedding space suitable for both label cleaning and the discovery of new activations. We then validate this method with an artificially corrupted label set based on source-separated high-density surface electromyography recordings, recovering the original timestamps even in extreme degrees of feature and class-dependent label noise. This approach enables a neural network to learn to accurately decode neurophysiological time series using any imperfect method of labelling the signal.

(0 min) The special and important problems of default prediction for municipal bonds are addressed using a combination of text embeddings from a pre-trained transformer network, a fully connected neural network, and synthetic oversampling. The combination of these techniques provides significant improvement in performance over human estimates, linear models, and boosted ensemble models, on data with extreme imbalance. Less than 0.2% of municipal bonds default, but our technique predicts 9 out of 10 defaults at the time of issue, without using bond ratings, at a cost of false positives on less than 0.1% non-defaulting bonds. The results hold the promise of reducing the cost of capital for local public goods, which are vital for society, and bring techniques previously used in personal credit and public equities (or national fixed income), as well as the current generation of embedding techniques, to sub-sovereign credit decisions.

(0 min) Continual Learning (CL) is an emerging machine learning paradigm that aims to learn from a continuous stream of tasks without forgetting knowledge learned from the previous tasks. To avoid performance decrease caused by forgetting, prior studies exploit episodic memory (EM), which stores a subset of the past observed samples while learning from new non-i.i.d. data. Despite the promising results, since CL is often assumed to execute on mobile or IoT devices, the EM size is bounded by the small hardware memory capacity and makes it infeasible to meet the accuracy requirements for real-world applications. Specifically, all prior CL methods discard samples overflowed from the EM and can never retrieve them back for subsequent training steps, incurring loss of information that would exacerbate catastrophic forgetting. We explore a novel hierarchical EM management strategy to address the forgetting issue. In particular, in mobile and IoT devices, real-time data can be stored not just in high-speed RAMs but in internal storage devices as well, which offer significantly larger capacity than the RAMs. Based on this insight, we propose to exploit the abundant storage to preserve past experiences and alleviate the forgetting by allowing CL to efficiently migrate samples between memory and storage without being interfered by the slow access speed of the storage. We call it Carousel Memory (CarM). As CarM is complementary to existing CL methods, we conduct extensive evaluations of our method with seven popular CL methods and show that CarM significantly improves the accuracy of the methods across different settings by large margins in final average accuracy (up to 28.4%) while retaining the same training efficiency.

(0 min) A novel signaling design for secure transmission over two-user multiple-input multiple-output non-orthogonal multiple access channel using deep neural networks (DNNs) is proposed. The goal of the DNN is to form the covariance matrix of users' signals such that the message of each user is transmitted reliably while being confidential from its counterpart. The proposed DNN linearly precodes each user's signal before superimposing them and achieves near-optimal performance with significantly lower run time. Simulation results show that the proposed models reach about 98% of the secrecy capacity rates. The spectral efficiency of the DNN precoder is much higher than that of existing analytical linear precoders--e.g., generalized singular value decomposition--and its on-the-fly complexity is several times less than the existing iterative methods.

(0 min) We consider language modelling (LM) as a multi-label structured prediction task by re-framing training from solely predicting a single ground-truth word to ranking a set of words which could continue a given context. To avoid annotating top-$k$ ranks, we generate them using pre-trained LMs: GPT-2, BERT, and Born-Again models. This leads to a rank-based form of knowledge distillation (KD). We also develop a method using $N$-grams to create a non-probabilistic teacher which generates the ranks without the need of a pre-trained LM. We confirm the hypotheses that we can treat LMing as a ranking task and that we can do so without the use of a pre-trained LM. We show that rank-based KD generally improves perplexity (PPL), often with statistical significance, when compared to Kullback-Leibler-based KD. Surprisingly, given the simplicity of the method, $N$-grams act as competitive teachers and achieve similar performance as using either BERT or a Born-Again model teachers. GPT-2 always acts as the best teacher, though, and using it and a Transformer-XL student on Wiki-02, rank-based KD reduces a cross-entropy baseline from 65.27 to 55.94 and against a KL-based KD of 56.70.

(0 min) Anomaly detection relies on designing a score to determine whether a particular event is uncharacteristic of a given background distribution. One way to define a score is to use autoencoders, which rely on the ability to reconstruct certain types of data (background) but not others (signals). In this paper, we study some challenges associated with variational autoencoders, such as the dependence on hyperparameters and the metric used, in the context of anomalous signal (top and $W$) jets in a QCD background. We find that the hyperparameter choices strongly affect the network performance and that the optimal parameters for one signal are non-optimal for another. In exploring the networks, we uncover a connection between the latent space of a variational autoencoder trained using mean-squared-error and the optimal transport distances within the dataset. We then show that optimal transport distances to representative events in the background dataset can be used directly for anomaly detection, with performance comparable to the autoencoders. Whether using autoencoders or optimal transport distances for anomaly detection, we find that the choices that best represent the background are not necessarily best for signal identification. These challenges with unsupervised anomaly detection bolster the case for additional exploration of semi-supervised or alternative approaches.

(0 min) We present and review an algorithmic and theoretical framework for improving neural network architecture design via momentum. As case studies, we consider how momentum can improve the architecture design for recurrent neural networks (RNNs), neural ordinary differential equations (ODEs), and transformers. We show that integrating momentum into neural network architectures has several remarkable theoretical and empirical benefits, including 1) integrating momentum into RNNs and neural ODEs can overcome the vanishing gradient issues in training RNNs and neural ODEs, resulting in effective learning long-term dependencies. 2) momentum in neural ODEs can reduce the stiffness of the ODE dynamics, which significantly enhances the computational efficiency in training and testing. 3) momentum can improve the efficiency and accuracy of transformers.

(0 min) The Possibilistic Fuzzy Local Information C-Means (PFLICM) method is presented as a technique to segment side-look synthetic aperture sonar (SAS) imagery into distinct regions of the sea-floor. In this work, we investigate and present the results of an automated feature selection approach for SAS image segmentation. The chosen features and resulting segmentation from the image will be assessed based on a select quantitative clustering validity criterion and the subset of the features that reach a desired threshold will be used for the segmentation process.

(0 min) We propose and assess an alternative quantum generator architecture in the context of generative adversarial learning for Monte Carlo event generation, used to simulate particle physics processes at the Large Hadron Collider (LHC). We validate this methodology by implementing the quantum network on artificial data generated from known underlying distributions. The network is then applied to Monte Carlo-generated datasets of specific LHC scattering processes. The new quantum generator architecture leads to an improvement in state-of-the-art implementations while maintaining shallow-depth networks. Moreover, the quantum generator successfully learns the underlying distribution functions even if trained with small training sample sets; this is particularly interesting for data augmentation applications. We deploy this novel methodology on two different quantum hardware architectures, trapped-ion and superconducting technologies, to test its hardware-independent viability.

(0 min) Bilevel optimization (BO) has arisen as a powerful tool for solving many modern machine learning problems. However, due to the nested structure of BO, existing gradient-based methods require second-order derivative approximations via Jacobian- or/and Hessian-vector computations, which can be very costly in practice, especially with large neural network models. In this work, we propose a novel BO algorithm, which adopts Evolution Strategies (ES) based method to approximate the response Jacobian matrix in the hypergradient of BO, and hence fully eliminates all second-order computations. We call our algorithm as ESJ (which stands for the ES-based Jacobian method) and further extend it to the stochastic setting as ESJ-S. Theoretically, we characterize the convergence guarantee and computational complexity for our algorithms. Experimentally, we demonstrate the superiority of our proposed algorithms compared to the state of the art methods on various bilevel problems. Particularly, in our experiment in the few-shot meta-learning problem, we meta-learn the twelve millions parameters of a ResNet-12 network over the miniImageNet dataset, which evidently demonstrates the scalability of our ES-based bilevel approach and its feasibility in the large-scale setting.

(0 min) Graph Convolutional Networks (GCN) with multi-hop aggregation is more expressive than one-hop GCN but suffers from higher model complexity. Finding the shortest aggregation range that achieves comparable expressiveness and minimizes this side effect remains an open question. We answer this question by showing that multi-layer second-order graph convolution (SoGC) is sufficient to attain the ability of expressing polynomial spectral filters with arbitrary coefficients. Compared to models with one-hop aggregation, multi-hop propagation, and jump connections, SoGC possesses filter representational completeness while being lightweight, efficient, and easy to implement. Thereby, we suggest that SoGC is a simple design capable of forming the basic building block of GCNs, playing the same role as $3 \times 3$ kernels in CNNs. We build our Second-Order Graph Convolutional Networks (SoGCN) with SoGC and design a synthetic dataset to verify its filter fitting capability to validate these points. For real-world tasks, we present the state-of-the-art performance of SoGCN on the benchmark of node classification, graph classification, and graph regression datasets.

(0 min) In this paper, we demonstrate how a generative model can be used to build a better recognizer through the control of content and style. We are building an online handwriting recognizer from a modest amount of training samples. By training our controllable handwriting synthesizer on the same data, we can synthesize handwriting with previously underrepresented content (e.g., URLs and email addresses) and style (e.g., cursive and slanted). Moreover, we propose a framework to analyze a recognizer that is trained with a mixture of real and synthetic training data. We use the framework to optimize data synthesis and demonstrate significant improvement on handwriting recognition over a model trained on real data only. Overall, we achieve a 66% reduction in Character Error Rate.

(0 min) Continuous blood pressure (BP) measurements can reflect a bodys response to diseases and serve as a predictor of cardiovascular and other health conditions. While current cuff-based BP measurement methods are incapable of providing continuous BP readings, invasive BP monitoring methods also tend to cause patient dissatisfaction and can potentially cause infection. In this research, we developed a method for estimating blood pressure based on the features extracted from Electrocardiogram (ECG) and Photoplethysmogram (PPG) signals and the Arterial Blood Pressure (ABP) data. The vector of features extracted from the preprocessed ECG and PPG signals is used in this approach, which include Pulse Transit Time (PTT), PPG Intensity Ratio (PIR), and Heart Rate (HR), as the input of a clustering algorithm and then developing separate regression models like Random Forest Regression, Gradient Boosting Regression, and Multilayer Perceptron Regression algorithms for each resulting cluster. We evaluated and compared the findings to create the model with the highest accuracy by applying the clustering approach and identifying the optimal number of clusters, and eventually the acceptable prediction model. The paper compares the results obtained with and without this clustering. The results show that the proposed clustering approach helps obtain more accurate estimates of Systolic Blood Pressure (SBP) and Diastolic Blood Pressure (DBP). Given the inconsistency, high dispersion, and multitude of trends in the datasets for different features, using the clustering approach improved the estimation accuracy by 50-60%.

(0 min) Recommender systems may be confounded by various types of confounding factors (also called confounders) that may lead to inaccurate recommendations and sacrificed recommendation performance. Current approaches to solving the problem usually design each specific model for each specific confounder. However, real-world systems may include a huge number of confounders and thus designing each specific model for each specific confounder is unrealistic. More importantly, except for those "explicit confounders" that researchers can manually identify and process such as item's position in the ranking list, there are also many "latent confounders" that are beyond the imagination of researchers. For example, users' rating on a song may depend on their current mood or the current weather, and users' preference on ice creams may depend on the air temperature. Such latent confounders may be unobservable in the recorded training data. To solve the problem, we propose a deconfounded causal collaborative filtering model. We first frame user behaviors with unobserved confounders into a causal graph, and then we design a front-door adjustment model carefully fused with machine learning to deconfound the influence of unobserved confounders. The proposed model is able to handle both global confounders and personalized confounders. Experiments on real-world e-commerce datasets show that our method is able to deconfound unobserved confounders to achieve better recommendation performance.

(0 min) Motivated by extreme multi-label classification applications, we consider training deep learning models over sparse data in multi-GPU servers. The variance in the number of non-zero features across training batches and the intrinsic GPU heterogeneity combine to limit accuracy and increase the time to convergence. We address these challenges with Adaptive SGD, an adaptive elastic model averaging stochastic gradient descent algorithm for heterogeneous multi-GPUs that is characterized by dynamic scheduling, adaptive batch size scaling, and normalized model merging. Instead of statically partitioning batches to GPUs, batches are routed based on the relative processing speed. Batch size scaling assigns larger batches to the faster GPUs and smaller batches to the slower ones, with the goal to arrive at a steady state in which all the GPUs perform the same number of model updates. Normalized model merging computes optimal weights for every GPU based on the assigned batches such that the combined model achieves better accuracy. We show experimentally that Adaptive SGD outperforms four state-of-the-art solutions in time-to-accuracy and is scalable with the number of GPUs.

(0 min) Click-through rate prediction (CTR) and post-click conversion rate prediction (CVR) play key roles across all industrial ranking systems, such as recommendation systems, online advertising, and search engines. Different from the extensive research on CTR, there is much less research on CVR estimation, whose main challenge is extreme data sparsity with one or two orders of magnitude reduction in the number of samples than CTR. People try to solve this problem with the paradigm of multi-task learning with the sufficient samples of CTR, but the typical hard sharing method can't effectively solve this problem, because it is difficult to analyze which parts of network components can be shared and which parts are in conflict, i.e., there is a large inaccuracy with artificially designed neurons sharing. In this paper, we model CVR in a brand-new method by adopting the lottery-ticket-hypothesis-based sparse sharing multi-task learning, which can automatically and flexibly learn which neuron weights to be shared without artificial experience. Experiments on the dataset gathered from traffic logs of Tencent video's recommendation system demonstrate that sparse sharing in the CVR model significantly outperforms competitive methods. Due to the nature of weight sparsity in sparse sharing, it can also significantly reduce computational complexity and memory usage which are very important in the industrial recommendation system.

(3 min) We define `ousiometrics' to be the study of essential meaning in whatever context that meaningful signals are communicated, and `telegnomics' as the study of remotely sensed knowledge. From work emerging through the middle of the 20th century, the essence of meaning has become generally accepted as being well captured by the three orthogonal dimensions of evaluation, potency, and activation (EPA). By re-examining first types and then tokens for the English language, and through the use of automatically annotated histograms -- `ousiograms' -- we find here that: 1. The essence of meaning conveyed by words is instead best described by a compass-like power-danger (PD) framework, and 2. Analysis of a disparate collection of large-scale English language corpora -- literature, news, Wikipedia, talk radio, and social media -- shows that natural language exhibits a systematic bias toward safe, low danger words -- a reinterpretation of the Pollyanna principle's positivity bias for written expression. To help justify our choice of dimension names and to help address the problems with representing observed ousiometric dimensions by bipolar adjective pairs, we introduce and explore `synousionyms' and `antousionyms' -- ousiometric counterparts of synonyms and antonyms. We further show that the PD framework revises the circumplex model of affect as a more general model of state of mind. Finally, we use our findings to construct and test a prototype `ousiometer', a telegnomic instrument that measures ousiometric time series for temporal corpora. We contend that our power-danger ousiometric framework provides a complement for entropy-based measurements, and may be of value for the study of a wide variety of communication across biological and artificial life.

(0 min) Dominant sentence ordering models can be classified into pairwise ordering models and set-to-sequence models. However, there is little attempt to combine these two types of models, which inituitively possess complementary advantages. In this paper, we propose a novel sentence ordering framework which introduces two classifiers to make better use of pairwise orderings for graph-based sentence ordering. Specially, given an initial sentence-entity graph, we first introduce a graph-based classifier to predict pairwise orderings between linked sentences. Then, in an iterative manner, based on the graph updated by previously predicted high-confident pairwise orderings, another classifier is used to predict the remaining uncertain pairwise orderings. At last, we adapt a GRN-based sentence ordering model on the basis of final graph. Experiments on five commonly-used datasets demonstrate the effectiveness and generality of our model. Particularly, when equipped with BERT and FHDecoder, our model achieves state-of-the-art performance.

(0 min) In a recent study of auditory evoked potential (AEP) based brain-computer interface (BCI), it was shown that, with an encoder-decoder framework, it is possible to translate human neural activity to speech (T-CAS). However, current encoder-decoder-based methods achieve T-CAS often with a two-step method where the information is passed between the encoder and decoder with a shared dimension reduction vector, which may result in a loss of information. A potential approach to this problem is to design an end-to-end method by using a dual generative adversarial network (DualGAN) without dimension reduction of passing information, but it cannot realize one-to-one signal-to-signal translation (see Fig.1 (a) and (b)). In this paper, we propose an end-to-end model to translate human neural activity to speech directly, create a new electroencephalogram (EEG) datasets for participants with good attention by design a device to detect participants' attention, and introduce a dual-dual generative adversarial network (Dual-DualGAN) (see Fig. 1 (c) and (d)) to address an end-to-end translation of human neural activity to speech (ET-CAS) problem by group labelling EEG signals and speech signals, inserting a transition domain to realize cross-domain mapping. In the transition domain, the transition signals are cascaded by the corresponding EEG and speech signals in a certain proportion, which can build bridges for EEG and speech signals without corresponding features, and realize one-to-one cross-domain EEG-to-speech translation. The proposed method can translate word-length and sentence-length sequences of neural activity to speech. Experimental evaluation has been conducted to show that the proposed method significantly outperforms state-of-the-art methods on both words and sentences of auditory stimulus.

(0 min) Event correlation reasoning infers whether a natural language paragraph containing multiple events conforms to human common sense. For example, "Andrew was very drowsy, so he took a long nap, and now he is very alert" is sound and reasonable. In contrast, "Andrew was very drowsy, so he stayed up a long time, now he is very alert" does not comply with human common sense. Such reasoning capability is essential for many downstream tasks, such as script reasoning, abductive reasoning, narrative incoherence, story cloze test, etc. However, conducting event correlation reasoning is challenging due to a lack of large amounts of diverse event-based knowledge and difficulty in capturing correlation among multiple events. In this paper, we propose EventBERT, a pre-trained model to encapsulate eventuality knowledge from unlabeled text. Specifically, we collect a large volume of training examples by identifying natural language paragraphs that describe multiple correlated events and further extracting event spans in an unsupervised manner. We then propose three novel event- and correlation-based learning objectives to pre-train an event correlation model on our created training corpus. Empirical results show EventBERT outperforms strong baselines on four downstream tasks, and achieves SoTA results on most of them. Besides, it outperforms existing pre-trained models by a large margin, e.g., 6.5~23%, in zero-shot learning of these tasks.

(0 min) The integration of syntactic structures into Transformer machine translation has shown positive results, but to our knowledge, no work has attempted to do so with semantic structures. In this work we propose two novel parameter-free methods for injecting semantic information into Transformers, both rely on semantics-aware masking of (some of) the attention heads. One such method operates on the encoder, through a Scene-Aware Self-Attention (SASA) head. Another on the decoder, through a Scene-Aware Cross-Attention (SACrA) head. We show a consistent improvement over the vanilla Transformer and syntax-aware models for four language pairs. We further show an additional gain when using both semantic and syntactic structures in some language pairs.

(3 min) Modern mathematics is built on the idea that proofs should be translatable into formal proofs, whose validity is an objective question, decidable by a computer. Yet, in practice, proofs are informal and may omit many details. An agent considers a proof valid if they trust that it could be expanded into a machine-verifiable proof. A proof's validity can thus become a subjective matter and lead to a debate, which may be difficult to settle. Hence, while the concept of valid proof is well-defined, the process to establish validity is itself a complex multi-agent problem. We introduce the SPRIG protocol. SPRIG allows agents to propose and verify succinct and informative proofs in a decentralized fashion; the trust is established by agents being able to request more details in the proof steps; debates, if they arise, must isolate details of proofs and, if they persist, go down to machine-level details, where they are automatically settled. A structure of bounties and stakes is set to incentivize agents to act in good faith. We propose a game-theoretic discussion of SPRIG, showing how agents with various types of information interact, leading to a proof tree with an appropriate level of detail and to the invalidation of wrong proofs, and we discuss resilience against various attacks. We then analyze a simplified model, characterize its equilibria and compute the agents' level of trust. SPRIG is designed to run as a smart contract on a blockchain platform. This allows anonymous agents to participate in the verification debate, and to contribute with their information. The smart contract mediates the interactions, settles debates, and guarantees that bounties and stakes are paid as specified. SPRIG enables new applications, such as the issuance of bounties for open problems, and the creation of derivatives markets, allowing agents to inject more information pertaining to proofs.

(0 min) Training machines to understand natural language and interact with humans is an elusive and essential task of artificial intelligence. A diversity of dialogue systems has been designed with the rapid development of deep learning techniques, especially the recent pre-trained language models (PrLMs). Among these studies, the fundamental yet challenging type of task is dialogue comprehension whose role is to teach the machines to read and comprehend the dialogue context before responding. In this paper, we review the previous methods from the technical perspective of dialogue modeling for the dialogue comprehension task. We summarize the characteristics and challenges of dialogue comprehension in contrast to plain-text reading comprehension. Then, we discuss three typical patterns of dialogue modeling. In addition, we categorize dialogue-related pre-training techniques which are employed to enhance PrLMs in dialogue scenarios. Finally, we highlight the technical advances in recent years and point out the lessons from the empirical analysis and the prospects towards a new frontier of researches.

(0 min) Pre-trained language models in the past years have shown exponential growth in model parameters and compute time. ELECTRA is a novel approach for improving the compute efficiency of pre-trained language models (e.g. BERT) based on masked language modeling (MLM) by addressing the sample inefficiency problem with the replaced token detection (RTD) task. Our work proposes adaptive early exit strategy to maximize the efficiency of the pre-training process by relieving the model's subsequent layers of the need to process latent features by leveraging earlier layer representations. Moreover, we evaluate an initial approach to the problem that has not succeeded in maintaining the accuracy of the model while showing a promising compute efficiency by thoroughly investigating the necessity of the generator module of ELECTRA.

(0 min) Automated essay scoring (AES) is gaining increasing attention in the education sector as it significantly reduces the burden of manual scoring and allows ad hoc feedback for learners. Natural language processing based on machine learning has been shown to be particularly suitable for text classification and AES. While many machine-learning approaches for AES still rely on a bag-of-words (BOW) approach, we consider a transformer-based approach in this paper, compare its performance to a logistic regression model based on the BOW approach and discuss their differences. The analysis is based on 2,088 email responses to a problem-solving task, that were manually labeled in terms of politeness. Both transformer models considered in that analysis outperformed without any hyper-parameter tuning the regression-based model. We argue that for AES tasks such as politeness classification, the transformer-based approach has significant advantages, while a BOW approach suffers from not taking word order into account and reducing the words to their stem. Further, we show how such models can help increase the accuracy of human raters, and we provide a detailed instruction on how to implement transformer-based models for one's own purpose.

(0 min) Negation is both an operation in formal logic and in natural language by which a proposition is replaced by one stating the opposite, as by the addition of "not" or another negation cue. Treating negation in an adequate way is required for cognitive reasoning, which aims at modeling the human ability to draw meaningful conclusions despite incomplete and inconsistent knowledge. One task of cognitive reasoning is answering questions given by sentences in natural language. There are tools based on discourse representation theory to convert sentences automatically into a formal logic representation, and additional knowledge can be added using the predicate names in the formula and knowledge databases. However, the knowledge in logic databases in practice always is incomplete. Hence, forward reasoning of automated reasoning systems alone does not suffice to derive answers to questions because, instead of complete proofs, often only partial positive knowledge can be derived, while negative knowledge is used only during the reasoning process. In consequence, we aim at eliminating syntactic negation, strictly speaking, the negated event or property. In this paper, we describe an effective procedure to determine the negated event or property in order to replace it by its inverse. This lays the basis of cognitive reasoning, employing both logic and machine learning for general question answering. We evaluate our procedure by several benchmarks and demonstrate its practical usefulness in our cognitive reasoning system.

(0 min) Time is an important dimension in our physical world. Lots of facts can evolve with respect to time. For example, the U.S. President might change every four years. Therefore, it is important to consider the time dimension and empower the existing QA models to reason over time. However, the existing QA datasets contain rather few time-sensitive questions, hence not suitable for diagnosing or benchmarking the model's temporal reasoning capability. In order to promote research in this direction, we propose to construct a time-sensitive QA dataset. The dataset is constructed by 1) mining time-evolving facts from WikiData and align them to their corresponding Wikipedia page, 2) employing crowd workers to verify and calibrate these noisy facts, 3) generating question-answer pairs based on the annotated time-sensitive facts. Our dataset poses challenges in the aspect of both temporal understanding and temporal reasoning. We evaluate different SoTA long-document QA systems like BigBird and FiD on our dataset. The best-performing model FiD can only achieve 46\% accuracy, still far behind the human performance of 87\%. We demonstrate that these models are still lacking the ability to perform consistent temporal reasoning. Therefore, we believe that our dataset could serve as a benchmark to develop NLP models more sensitive to temporal shift. The dataset and code are released in~\url{https://github.com/wenhuchen/Time-Sensitive-QA}.

(0 min) Existing generative pre-trained language models (e.g., GPT) focus on modeling the language structure and semantics of general texts. However, those models do not consider the numerical properties of numbers and cannot perform robustly on numerical reasoning tasks (e.g., math word problems and measurement estimation). In this paper, we propose NumGPT, a generative pre-trained model that explicitly models the numerical properties of numbers in texts. Specifically, it leverages a prototype-based numeral embedding to encode the mantissa of the number and an individual embedding to encode the exponent of the number. A numeral-aware loss function is designed to integrate numerals into the pre-training objective of NumGPT. We conduct extensive experiments on four different datasets to evaluate the numeracy ability of NumGPT. The experiment results show that NumGPT outperforms baseline models (e.g., GPT and GPT with DICE) on a range of numerical reasoning tasks such as measurement estimation, number comparison, math word problems, and magnitude classification. Ablation studies are also conducted to evaluate the impact of pre-training and model hyperparameters on the performance.

(0 min) Zero/few-shot transfer to unseen services is a critical challenge in task-oriented dialogue research. The Schema-Guided Dialogue (SGD) dataset introduced a paradigm for enabling models to support an unlimited number of services without additional data collection or re-training through the use of schemas. Schemas describe service APIs in natural language, which models consume to understand the services they need to support. However, the impact of the choice of language in these schemas on model performance remains unexplored. We address this by releasing SGD-X, a benchmark for measuring the robustness of dialogue systems to linguistic variations in schemas. SGD-X extends the SGD dataset with crowdsourced variants for every schema, where variants are semantically similar yet stylistically diverse. We evaluate two dialogue state tracking models on SGD-X and observe that neither generalizes well across schema variations, measured by joint goal accuracy and a novel metric for measuring schema sensitivity. Furthermore, we present a simple model-agnostic data augmentation method to improve schema robustness and zero-shot generalization to unseen services.

(0 min) Despite their recent popularity and well known advantages, dense retrievers still lag behind sparse methods such as BM25 in their ability to reliably match salient phrases and rare entities in the query. It has been argued that this is an inherent limitation of dense models. We disprove this claim by introducing the Salient Phrase Aware Retriever (SPAR), a dense retriever with the lexical matching capacity of a sparse model. In particular, we show that a dense retriever {\Lambda} can be trained to imitate a sparse one, and SPAR is built by augmenting a standard dense retriever with {\Lambda}. When evaluated on five open-domain question answering datasets and the MS MARCO passage retrieval task, SPAR sets a new state of the art for dense and sparse retrievers and can match or exceed the performance of more complicated dense-sparse hybrid systems.

(0 min) Variation in speech is often represented and investigated using phonetic transcriptions, but transcribing speech is time-consuming and error prone. As an alternative representation, therefore, we investigate the extraction of acoustic embeddings from several self-supervised neural models. We use these representations to compute word-based pronunciation differences between non-native and native speakers of English, and between different dialect pronunciations, and evaluate these differences by comparing them with available human native-likeness judgments. We show that Transformer-based speech representations lead to significant performance gains over the use of phonetic transcriptions, and find that feature-based use of Transformer models is most effective with one of the middle layers instead of the final layer. We also demonstrate that these neural speech representations not only capture segmental differences, but also intonational and durational differences that cannot be represented by a set of discrete symbols used in phonetic transcriptions.

(0 min) The evaluation of question answering models compares ground-truth annotations with model predictions. However, as of today, this comparison is mostly lexical-based and therefore misses out on answers that have no lexical overlap but are still semantically similar, thus treating correct answers as false. This underestimation of the true performance of models hinders user acceptance in applications and complicates a fair comparison of different models. Therefore, there is a need for an evaluation metric that is based on semantics instead of pure string similarity. In this short paper, we present SAS, a cross-encoder-based metric for the estimation of semantic answer similarity, and compare it to seven existing metrics. To this end, we create an English and a German three-way annotated evaluation dataset containing pairs of answers along with human judgment of their semantic similarity, which we release along with an implementation of the SAS metric and the experiments. We find that semantic similarity metrics based on recent transformer models correlate much better with human judgment than traditional lexical similarity metrics on our two newly created datasets and one dataset from related work.

(0 min) Although pre-trained models (PLMs) have achieved remarkable improvements in a wide range of NLP tasks, they are expensive in terms of time and resources. This calls for the study of training more efficient models with less computation but still ensures impressive performance. Instead of pursuing a larger scale, we are committed to developing lightweight yet more powerful models trained with equal or less computation and friendly to rapid deployment. This technical report releases our pre-trained model called Mengzi, which stands for a family of discriminative, generative, domain-specific, and multimodal pre-trained model variants, capable of a wide range of language and vision tasks. Compared with public Chinese PLMs, Mengzi is simple but more powerful. Our lightweight model has achieved new state-of-the-art results on the widely-used CLUE benchmark with our optimized pre-training and fine-tuning techniques. Without modifying the model architecture, our model can be easily employed as an alternative to existing PLMs. Our sources are available at https://github.com/Langboat/Mengzi.

(0 min) This paper presents a novel open-domain dialogue generation framework emphasizing the differentiation of speakers in multi-turn conversations. Differing from prior work that solely relies on the content of conversation history to generate a response, we argue that capturing relative social relations among utterances (i.e., generated by either the same speaker or different persons) benefits the machine capturing fine-grained context information from a conversation history to improve context coherence in the generated response. Given that, we propose a speaker-aware framework, named Parallel Hierarchical Attentive Encoder-Decoder (PHAED), that aims to model each utterance with the awareness of its speaker and contextual associations with the same speaker's previous messages. Specifically, in a conversation involving two speakers, we regard the utterances from one speaker as responses and those from the other as queries. After understanding queries via our encoder with inner-query and inter-query encodings, our decoder reuses the hidden states of previously generated responses to generate a new response. Our empirical results show that PHAED outperforms the state-of-the-art in both automatic and human evaluations. Furthermore, our ablation study shows that dialogue models with speaker tokens can generally decrease the possibility of generating non-coherent responses regarding the conversation context.

(0 min) We describe a Question Answering (QA) dataset that contains complex questions with conditional answers, i.e. the answers are only applicable when certain conditions apply. We call this dataset ConditionalQA. In addition to conditional answers, the dataset also features: (1) long context documents with information that is related in logically complex ways; (2) multi-hop questions that require compositional logical reasoning; (3) a combination of extractive questions, yes/no questions, questions with multiple answers, and not-answerable questions; (4) questions asked without knowing the answers. We show that ConditionalQA is challenging for many of the existing QA models, especially in selecting answer conditions. We believe that this dataset will motivate further research in answering complex questions over long documents. Data and leaderboard are publicly available at \url{https://github.com/haitian-sun/ConditionalQA}.

(2 min) The outbreak of the SARS-CoV-2 pandemic of the new COVID-19 disease (COVID-19 for short) demands empowering existing medical, economic, and social emergency backend systems with data analytics capabilities. An impediment in taking advantages of data analytics in these systems is the lack of a unified framework or reference model. Ontologies are highlighted as a promising solution to bridge this gap by providing a formal representation of COVID-19 concepts such as symptoms, infections rate, contact tracing, and drug modelling. Ontology-based solutions enable the integration of diverse data sources that leads to a better understanding of pandemic data, management of smart lockdowns by identifying pandemic hotspots, and knowledge-driven inference, reasoning, and recommendations to tackle surrounding issues.

(0 min) The mismatch between an external language model (LM) and the implicitly learned internal LM (ILM) of RNN-Transducer (RNN-T) can limit the performance of LM integration such as simple shallow fusion. A Bayesian interpretation suggests to remove this sequence prior as ILM correction. In this work, we study various ILM correction-based LM integration methods formulated in a common RNN-T framework. We provide a decoding interpretation on two major reasons for performance improvement with ILM correction, which is further experimentally verified with detailed analysis. We also propose an exact-ILM training framework by extending the proof given in the hybrid autoregressive transducer, which enables a theoretical justification for other ILM approaches. Systematic comparison is conducted for both in-domain and cross-domain evaluation on the Librispeech and TED-LIUM Release 2 corpora, respectively. Our proposed exact-ILM training can further improve the best ILM method.

(0 min) The Winograd Schema (WS) has been proposed as a test for measuring commonsense capabilities of models. Recently, pre-trained language model-based approaches have boosted performance on some WS benchmarks but the source of improvement is still not clear. This paper suggests that the apparent progress on WS may not necessarily reflect progress in commonsense reasoning. To support this claim, we first show that the current evaluation method of WS is sub-optimal and propose a modification that uses twin sentences for evaluation. We also propose two new baselines that indicate the existence of artifacts in WS benchmarks. We then develop a method for evaluating WS-like sentences in a zero-shot setting to account for the commonsense reasoning abilities acquired during the pretraining and observe that popular language models perform randomly in this setting when using our more strict evaluation. We conclude that the observed progress is mostly due to the use of supervision in training WS models, which is not likely to successfully support all the required commonsense reasoning skills and knowledge.

(0 min) Keyphrases, that concisely summarize the high-level topics discussed in a document, can be categorized into present keyphrase which explicitly appears in the source text, and absent keyphrase which does not match any contiguous subsequence but is highly semantically related to the source. Most existing keyphrase generation approaches synchronously generate present and absent keyphrases without explicitly distinguishing these two categories. In this paper, a Select-Guide-Generate (SGG) approach is proposed to deal with present and absent keyphrase generation separately with different mechanisms. Specifically, SGG is a hierarchical neural network which consists of a pointing-based selector at low layer concentrated on present keyphrase generation, a selection-guided generator at high layer dedicated to absent keyphrase generation, and a guider in the middle to transfer information from selector to generator. Experimental results on four keyphrase generation benchmarks demonstrate the effectiveness of our model, which significantly outperforms the strong baselines for both present and absent keyphrases generation. Furthermore, we extend SGG to a title generation task which indicates its extensibility in natural language generation tasks.

(0 min) In many contexts, lying -- the use of verbal falsehoods to deceive -- is harmful. While lying has traditionally been a human affair, AI systems that make sophisticated verbal statements are becoming increasingly prevalent. This raises the question of how we should limit the harm caused by AI "lies" (i.e. falsehoods that are actively selected for). Human truthfulness is governed by social norms and by laws (against defamation, perjury, and fraud). Differences between AI and humans present an opportunity to have more precise standards of truthfulness for AI, and to have these standards rise over time. This could provide significant benefits to public epistemics and the economy, and mitigate risks of worst-case AI futures. Establishing norms or laws of AI truthfulness will require significant work to: (1) identify clear truthfulness standards; (2) create institutions that can judge adherence to those standards; and (3) develop AI systems that are robustly truthful. Our initial proposals for these areas include: (1) a standard of avoiding "negligent falsehoods" (a generalisation of lies that is easier to assess); (2) institutions to evaluate AI systems before and after real-world deployment; and (3) explicitly training AI systems to be truthful via curated datasets and human interaction. A concerning possibility is that evaluation mechanisms for eventual truthfulness standards could be captured by political interests, leading to harmful censorship and propaganda. Avoiding this might take careful attention. And since the scale of AI speech acts might grow dramatically over the coming decades, early truthfulness standards might be particularly important because of the precedents they set.

(0 min) Question generation is a conditioned language generation task that consists in generating a context-aware question given a context and the targeted answer. Train language modelling with a mere likelihood maximization has been widely used while suffering from exposure bias and the discordance between the training and the test metrics. In the way of addressing this issue, The presented work portrays a fully Transformer-based reinforcement learning generator-evaluation architecture for neural question generation. To edge the flexibility of the generation, a semantic-based reward score was externally infused during the training to drive the training of the language model. The global architecture is laid out in a generator-evaluator fashion optimized directly to n-gram and semantic-based metrics. Evaluation metrics for language modelling only based on n-gram overlapping do not consider semantic relations between reference and candidate sequences. To improve the evaluation step, a two-fold evaluation was carried out. On the one side, an n-gram overlapping evaluation using the BLEU score. On the other side, a semantic-based assessment using BERTScore and NUBIA. The results were corroborated by a binary human evaluation of the semantic relatedness of the generated question and the ground truth. The results obtained showed that use a semantic-based REINFORCE algorithm for the question generation syntactically reshapes the generated questions while preserving their underlying semantic meaning. Many downstream applications can be drawn from a successful question generation including the enlargement of question answering datasets, the improvement of conversational systems, the enhancement of autonomous educational assessment systems, and so forth.

(2 min) In this paper, we discuss the initial attempts at boosting understanding human language based on deep-learning models with quantum computing. We successfully train a quantum-enhanced Long Short-Term Memory network to perform the parts-of-speech tagging task via numerical simulations. Moreover, a quantum-enhanced Transformer is proposed to perform the sentiment analysis based on the existing dataset.

(2 min) The NLP pipeline has evolved dramatically in the last few years. The first step in the pipeline is to find suitable annotated datasets to evaluate the tasks we are trying to solve. Unfortunately, most of the published datasets lack metadata annotations that describe their attributes. Not to mention, the absence of a public catalogue that indexes all the publicly available datasets related to specific regions or languages. When we consider low-resource dialectical languages, for example, this issue becomes more prominent. In this paper we create \textit{Masader}, the largest public catalogue for Arabic NLP datasets, which consists of 200 datasets annotated with 25 attributes. Furthermore, We develop a metadata annotation strategy that could be extended to other languages. We also make remarks and highlight some issues about the current status of Arabic NLP datasets and suggest recommendations to address them.

(2 min) Neural conversational models have long suffered from the problem of inconsistency and lacking coherent personality. To address the issue, persona-based models capturing individual characteristics have been proposed, but they still face the dilemma of model adaption and data privacy. To break this dilemma, we propose a novel Federated Natural Language Generation (FedNLG) framework, which learns personalized representations from various dataset on distributed devices, and thus implements the personalized dialogue system efficiently and safely. FedNLG first pre-trains parameters of standard neural conversational model over a large dialogue corpus, and then fine-tune the model parameters and persona embeddings on specific datasets, in a federated manner. Thus, the model could simultaneously learn the persona embeddings in local clients and learn shared model parameters by federated aggregation, which achieves accuracyprivacy balance. By conducting extensive experiments, we demonstrate the effectiveness of our model by pre-training model over Cornell Movie-Dialogs Corpus and fine-tuning the model over two TV series dataset.

(2 min) Semantic role labeling is a fundamental yet challenging task in the NLP community. Recent works of SRL mainly fall into two lines:1) BIO-based and 2) span-based. Despite effectiveness, they share some intrinsic drawbacks of not explicitly considering internal argument structures, which may potentially hinder the model's expressiveness. To remedy this, we propose to reduce SRL to a dependency parsing task and regard the flat argument spans as latent subtrees. In particular, we equip our formulation with a novel span-constrained TreeCRF model to make tree structures span-aware, and further extend it to the second-order case. Experiments on CoNLL05 and CoNLL12 benchmarks reveal that the results of our methods outperform all previous works and achieve the state-of-the-art.

(3 min) Human-AI collaborative policy synthesis is a procedure in which (1) a human initializes an autonomous agent's behavior, (2) Reinforcement Learning improves the human specified behavior, and (3) the agent can explain the final optimized policy to the user. This paradigm leverages human expertise and facilitates a greater insight into the learned behaviors of an agent. Existing approaches to enabling collaborative policy specification involve black box methods which are unintelligible and are not catered towards non-expert end-users. In this paper, we develop a novel collaborative framework to enable humans to initialize and interpret an autonomous agent's behavior, rooted in principles of human-centered design. Through our framework, we enable humans to specify an initial behavior model in the form of unstructured, natural language, which we then convert to lexical decision trees. Next, we are able to leverage these human-specified policies, to warm-start reinforcement learning and further allow the agent to optimize the policies through reinforcement learning. Finally, to close the loop on human-specification, we produce explanations of the final learned policy, in multiple modalities, to provide the user a final depiction about the learned policy of the agent. We validate our approach by showing that our model can produce >80% accuracy, and that human-initialized policies are able to successfully warm-start RL. We then conduct a novel human-subjects study quantifying the relative subjective and objective benefits of varying XAI modalities(e.g., Tree, Language, and Program) for explaining learned policies to end-users, in terms of usability and interpretability and identify the circumstances that influence these measures. Our findings emphasize the need for personalized explainable systems that can facilitate user-centric policy explanations for a variety of end-users.

(2 min) The ability to generate natural-language questions with controlled complexity levels is highly desirable as it further expands the applicability of question generation. In this paper, we propose an end-to-end neural complexity-controllable question generation model, which incorporates a mixture of experts (MoE) as the selector of soft templates to improve the accuracy of complexity control and the quality of generated questions. The soft templates capture question similarity while avoiding the expensive construction of actual templates. Our method introduces a novel, cross-domain complexity estimator to assess the complexity of a question, taking into account the passage, the question, the answer and their interactions. The experimental results on two benchmark QA datasets demonstrate that our QG model is superior to state-of-the-art methods in both automatic and manual evaluation. Moreover, our complexity estimator is significantly more accurate than the baselines in both in-domain and out-domain settings.

(2 min) Pre-trained language models (PLMs) aim to learn universal language representations by conducting self-supervised training tasks on large-scale corpora. Since PLMs capture word semantics in different contexts, the quality of word representations highly depends on word frequency, which usually follows a heavy-tailed distributions in the pre-training corpus. Therefore, the embeddings of rare words on the tail are usually poorly optimized. In this work, we focus on enhancing language model pre-training by leveraging definitions of the rare words in dictionaries (e.g., Wiktionary). To incorporate a rare word definition as a part of input, we fetch its definition from the dictionary and append it to the end of the input text sequence. In addition to training with the masked language modeling objective, we propose two novel self-supervised pre-training tasks on word and sentence-level alignment between input text sequence and rare word definitions to enhance language modeling representation with dictionary. We evaluate the proposed Dict-BERT model on the language understanding benchmark GLUE and eight specialized domain benchmark datasets. Extensive experiments demonstrate that Dict-BERT can significantly improve the understanding of rare words and boost model performance on various NLP downstream tasks.

(2 min) Recent studies in speech perception have been closely linked to fields of cognitive psychology, phonology, and phonetics in linguistics. During perceptual attunement, a critical and sensitive developmental trajectory has been examined in bilingual and monolingual infants where they can best discriminate common phonemes. In this paper, we compare and identify these cognitive aspects on deep neural-based visual lip-reading models. We conduct experiments on the two most extensive public visual speech recognition datasets for English and Mandarin. Through our experimental results, we observe a strong correlation between these theories in cognitive psychology and our unique modeling. We inspect how these computational models develop similar phases in speech perception and acquisitions.

(2 min) Recent research on dialogue response selection has been mainly focused on selecting a proper response from a pre-defined small set of candidates using sophisticated neural models. Due to their heavy computational overhead, they are unable to select responses from a large candidate pool. In this study, we present a solution to directly select proper responses from a large corpus or even a nonparallel corpus that only consists of unpaired sentences, using a dense retrieval model. We extensively test our proposed approach under two experiment settings: (i) re-rank experiment that aims to rank a small set of pre-defined candidates; (ii) full-rank experiment where the target is to directly select proper responses from a full candidate pool that may contain millions of candidates. For re-rank setting, the superiority is quite surprising given its simplicity. For full-rank setting, we can emphasize that we are the first to do such evaluation. Moreover, human evaluation results show that increasing the size of nonparallel corpus leads to further improvement of our model performance\footnote{All our source codes, models and other related resources are publically available at \url{https://github.com/gmftbyGMFTBY/SimpleReDial-v1}.

(2 min) Compositionality, or the ability to combine familiar units like words into novel phrases and sentences, has been the focus of intense interest in artificial intelligence in recent years. To test compositional generalization in semantic parsing, Keysers et al. (2020) introduced Compositional Freebase Queries (CFQ). This dataset maximizes the similarity between the test and train distributions over primitive units, like words, while maximizing the compound divergence: the dissimilarity between test and train distributions over larger structures, like phrases. Dependency parsing, however, lacks a compositional generalization benchmark. In this work, we introduce a gold-standard set of dependency parses for CFQ, and use this to analyze the behavior of a state-of-the art dependency parser (Qi et al., 2020) on the CFQ dataset. We find that increasing compound divergence degrades dependency parsing performance, although not as dramatically as semantic parsing performance. Additionally, we find the performance of the dependency parser does not uniformly degrade relative to compound divergence, and the parser performs differently on different splits with the same compound divergence. We explore a number of hypotheses for what causes the non-uniform degradation in dependency parsing performance, and identify a number of syntactic structures that drive the dependency parser's lower performance on the most challenging splits.

(2 min) To capture the semantic graph structure from raw text, most existing summarization approaches are built on GNNs with a pre-trained model. However, these methods suffer from cumbersome procedures and inefficient computations for long-text documents. To mitigate these issues, this paper proposes HETFORMER, a Transformer-based pre-trained model with multi-granularity sparse attentions for long-text extractive summarization. Specifically, we model different types of semantic nodes in raw text as a potential heterogeneous graph and directly learn heterogeneous relationships (edges) among nodes by Transformer. Extensive experiments on both single- and multi-document summarization tasks show that HETFORMER achieves state-of-the-art performance in Rouge F1 while using less memory and fewer parameters.

(2 min) Automatic Speech Recognition (ASR) systems have found their use in numerous industrial applications in very diverse domains. Since domain-specific systems perform better than their generic counterparts on in-domain evaluation, the need for memory and compute-efficient domain adaptation is obvious. Particularly, adapting parameter-heavy transformer-based language models used for rescoring ASR hypothesis is challenging. In this work, we overcome the problem using prompt-tuning, a methodology that trains a small number of domain token embedding parameters to prime a transformer-based LM to a particular domain. With just a handful of extra parameters per domain, we achieve much better perplexity scores over the baseline of using an unadapted LM. Despite being parameter-efficient, these improvements are comparable to those of fully-fine-tuned models with hundreds of millions of parameters. We replicate our findings in perplexity numbers to Word Error Rate in a domain-specific ASR system for one such domain.

(2 min) Improving the quality of Natural Language Understanding (NLU) models, and more specifically, task-oriented semantic parsing models, in production is a cumbersome task. In this work, we present a system called AutoNLU, which we designed to scale the NLU quality improvement process. It adds automation to three key steps: detection, attribution, and correction of model errors, i.e., bugs. We detected four times more failed tasks than with random sampling, finding that even a simple active learning sampling method on an uncalibrated model is surprisingly effective for this purpose. The AutoNLU tool empowered linguists to fix ten times more semantic parsing bugs than with prior manual processes, auto-correcting 65% of all identified bugs.

(2 min) In previous work, we have proposed the Audio-Visual Scene-Aware Dialog (AVSD) task, collected an AVSD dataset, developed AVSD technologies, and hosted an AVSD challenge track at both the 7th and 8th Dialog System Technology Challenges (DSTC7, DSTC8). In these challenges, the best-performing systems relied heavily on human-generated descriptions of the video content, which were available in the datasets but would be unavailable in real-world applications. To promote further advancements for real-world applications, we proposed a third AVSD challenge, at DSTC10, with two modifications: 1) the human-created description is unavailable at inference time, and 2) systems must demonstrate temporal reasoning by finding evidence from the video to support each answer. This paper introduces the new task that includes temporal reasoning and our new extension of the AVSD dataset for DSTC10, for which we collected human-generated temporal reasoning data. We also introduce a baseline system built using an AV-transformer, which we released along with the new dataset. Finally, this paper introduces a new system that extends our baseline system with attentional multimodal fusion, joint student-teacher learning (JSTL), and model combination techniques, achieving state-of-the-art performances on the AVSD datasets for DSTC7, DSTC8, and DSTC10. We also propose two temporal reasoning methods for AVSD: one attention-based, and one based on a time-domain region proposal network.

(2 min) We demonstrate that large language models are able to simulate Task Oriented Dialogues in novel domains, provided only with an API implementation and a list of goals. We show these simulations can formulate online, automatic metrics that correlate well with human evaluations. Furthermore, by checking for whether the User's goals are met, we can use simulation to repeatedly generate training data and improve the quality of simulations themselves. With no human intervention or domain-specific training data, our simulations bootstrap end-to-end models which achieve a 37\% error reduction in previously unseen domains. By including as few as 32 domain-specific conversations, bootstrapped models can match the performance of a fully-supervised model with $10\times$ more data. To our knowledge, this is the first time simulations have been shown to be effective at bootstrapping models without explicitly requiring any domain-specific training data, rule-engineering, or humans-in-the-loop.

(2 min) This paper addresses the problem of fake news detection in Spanish using Machine Learning techniques. It is fundamentally the same problem tackled for the English language; however, there is not a significant amount of publicly available and adequately labeled fake news in Spanish to effectively train a Machine Learning model, similarly to those proposed for the English language. Therefore, this work explores different training strategies and architectures to establish a baseline for further research in this area. Four datasets were used, two in English and two in Spanish, and four experimental schemes were tested, including a baseline with classical Machine Learning models, trained and validated using a small dataset in Spanish. The remaining schemes include state-of-the-art Deep Learning models trained (or fine-tuned) and validated in English, trained and validated in Spanish, and fitted in English and validated with automatic translated Spanish sentences. The Deep Learning architectures were built on top of different pre-trained Word Embedding representations, including GloVe, ELMo, BERT, and BETO (a BERT version trained on a large corpus in Spanish). According to the results, the best strategy was a combination of a pre-trained BETO model and a Recurrent Neural Network based on LSTM layers, yielding an accuracy of up to 80%; nonetheless, a baseline model using a Random Forest estimator obtained similar outcomes. Additionally, the translation strategy did not yield acceptable results because of the propagation error; there was also observed a significant difference in models performance when trained in English or Spanish, mainly attributable to the number of samples available for each language.

(2 min) Keyphrases are phrases in a document providing a concise summary of core content, helping readers to understand what the article is talking about in a minute. However, existing unsupervised works are not robust enough to handle various types of documents owing to the mismatch of sequence length for comparison. In this paper, we propose a novel unsupervised keyword extraction method by leveraging the BERT-based model to select and rank candidate keyphrases with a MASK strategy. In addition, we further enhance the model, denoted as Keyphrases Extraction BERT (KPEBERT), via designing a compatible self-supervised task and conducting a contrast learning. We conducted extensive experimental evaluation to demonstrate the superiority and robustness of the proposed method as well as the effectiveness of KPEBERT.

(2 min) We give simpler, sparser, and faster algorithms for differentially private fine-tuning of large-scale pre-trained language models, which achieve the state-of-the-art privacy versus utility tradeoffs on many standard NLP tasks. We propose a meta-framework for this problem, inspired by the recent success of highly parameter-efficient methods for fine-tuning. Our experiments show that differentially private adaptations of these approaches outperform previous private algorithms in three important dimensions: utility, privacy, and the computational and memory cost of private training. On many commonly studied datasets, the utility of private models approaches that of non-private models. For example, on the MNLI dataset we achieve an accuracy of $87.8\%$ using RoBERTa-Large and $83.5\%$ using RoBERTa-Base with a privacy budget of $\epsilon = 6.7$. In comparison, absent privacy constraints, RoBERTa-Large achieves an accuracy of $90.2\%$. Our findings are similar for natural language generation tasks. Privately fine-tuning with DART, GPT-2-Small, GPT-2-Medium, GPT-2-Large, and GPT-2-XL achieve BLEU scores of 38.5, 42.0, 43.1, and 43.8 respectively (privacy budget of $\epsilon = 6.8,\delta=$ 1e-5) whereas the non-private baseline is $48.1$. All our experiments suggest that larger models are better suited for private fine-tuning: while they are well known to achieve superior accuracy non-privately, we find that they also better maintain their accuracy when privacy is introduced.

(2 min) Computational modeling of the emotions evoked by art in humans is a challenging problem because of the subjective and nuanced nature of art and affective signals. In this paper, we consider the above-mentioned problem of understanding emotions evoked in viewers by artwork using both text and visual modalities. Specifically, we analyze images and the accompanying text captions from the viewers expressing emotions as a multimodal classification task. Our results show that single-stream multimodal transformer-based models like MMBT and VisualBERT perform better compared to both image-only models and dual-stream multimodal models having separate pathways for text and image modalities. We also observe improvements in performance for extreme positive and negative emotion classes, when a single-stream model like MMBT is compared with a text-only transformer model like BERT.

(2 min) Pre-trained language models have recently been shown to be able to perform translation without finetuning via prompting. Inspired by these findings, we study improving the performance of pre-trained language models on translation tasks, where training neural machine translation models is the current de facto approach. We present Multi-Stage Prompting, a simple and lightweight approach for better adapting pre-trained language models to translation tasks. To make pre-trained language models better translators, we divide the translation process via pre-trained language models into three separate stages: the encoding stage, the re-encoding stage, and the decoding stage. During each stage, we independently apply different continuous prompts for allowing pre-trained language models better adapting to translation tasks. We conduct extensive experiments on low-, medium-, and high-resource translation tasks. Experiments show that our method can significantly improve the translation performance of pre-trained language models.

(2 min) The COVID-19 pandemic poses a great threat to global public health. Meanwhile, there is massive misinformation associated with the pandemic which advocates unfounded or unscientific claims. Even major social media and news outlets have made an extra effort in debunking COVID-19 misinformation, most of the fact-checking information is in English, whereas some unmoderated COVID-19 misinformation is still circulating in other languages, threatening the health of less-informed people in immigrant communities and developing countries. In this paper, we make the first attempt to detect COVID-19 misinformation in a low-resource language (Chinese) only using the fact-checked news in a high-resource language (English). We start by curating a Chinese real&fake news dataset according to existing fact-checking information. Then, we propose a deep learning framework named CrossFake to jointly encode the cross-lingual news body texts and capture the news content as much as possible. Empirical results on our dataset demonstrate the effectiveness of CorssFake under the cross-lingual setting and it also outperforms several monolingual and cross-lingual fake news detectors. The dataset is available at https://github.com/YingtongDou/CrossFake.

(2 min) Natural language processing (NLP) systems have become a central technology in communication, education, medicine, artificial intelligence, and many other domains of research and development. While the performance of NLP methods has grown enormously over the last decade, this progress has been restricted to a minuscule subset of the world's 6,500 languages. We introduce a framework for estimating the global utility of language technologies as revealed in a comprehensive snapshot of recent publications in NLP. Our analyses involve the field at large, but also more in-depth studies on both user-facing technologies (machine translation, language understanding, question answering, text-to-speech synthesis) as well as more linguistic NLP tasks (dependency parsing, morphological inflection). In the process, we (1) quantify disparities in the current state of NLP research, (2) explore some of its associated societal and academic factors, and (3) produce tailored recommendations for evidence-based policy making aimed at promoting more global and equitable language technologies.

(2 min) Entity Alignment (EA) aims to match equivalent entities across different Knowledge Graphs (KGs) and is an essential step of KG fusion. Current mainstream methods -- neural EA models -- rely on training with seed alignment, i.e., a set of pre-aligned entity pairs which are very costly to annotate. In this paper, we devise a novel Active Learning (AL) framework for neural EA, aiming to create highly informative seed alignment to obtain more effective EA models with less annotation cost. Our framework tackles two main challenges encountered when applying AL to EA: (1) How to exploit dependencies between entities within the AL strategy. Most AL strategies assume that the data instances to sample are independent and identically distributed. However, entities in KGs are related. To address this challenge, we propose a structure-aware uncertainty sampling strategy that can measure the uncertainty of each entity as well as its impact on its neighbour entities in the KG. (2) How to recognise entities that appear in one KG but not in the other KG (i.e., bachelors). Identifying bachelors would likely save annotation budget. To address this challenge, we devise a bachelor recognizer paying attention to alleviate the effect of sampling bias. Empirical results show that our proposed AL strategy can significantly improve sampling quality with good generality across different datasets, EA models and amount of bachelors.

(2 min) Dialogue agents that interact with humans in situated environments need to manage referential ambiguity across multiple modalities and ask for help as needed. However, it is not clear what kinds of questions such agents should ask nor how the answers to such questions can be used to resolve ambiguity. To address this, we analyzed dialogue data from an interactive study in which participants controlled a virtual robot tasked with organizing a set of tools while engaging in dialogue with a live, remote experimenter. We discovered a number of novel results, including the distribution of question types used to resolve ambiguity and the influence of dialogue-level factors on the reference resolution process. Based on these empirical findings we: (1) developed a computational model for clarification requests using a decision network with an entropy-based utility assignment method that operates across modalities, (2) evaluated the model, showing that it outperforms a slot-filling baseline in environments of varying ambiguity, and (3) interpreted the results to offer insight into the ways that agents can ask questions to facilitate situated reference resolution.

(2 min) Pairwise dot product-based attention allows Transformers to exchange information between tokens in an input-dependent way, and is key to their success across diverse applications in language and vision. However, a typical Transformer model computes such pairwise attention scores repeatedly for the same sequence, in multiple heads in multiple layers. We systematically analyze the empirical similarity of these scores across heads and layers and find them to be considerably redundant, especially adjacent layers showing high similarity. Motivated by these findings, we propose a novel architecture that reuses attention scores computed in one layer in multiple subsequent layers. Experiments on a number of standard benchmarks show that reusing attention delivers performance equivalent to or better than standard transformers, while reducing both compute and memory usage.

(2 min) We propose a novel method for applying Transformer models to extractive question answering (QA) tasks. Recently, pretrained generative sequence-to-sequence (seq2seq) models have achieved great success in question answering. Contributing to the success of these models are internal attention mechanisms such as cross-attention. We propose a simple strategy to obtain an extractive answer span from the generative model by leveraging the decoder cross-attention patterns. Viewing cross-attention as an architectural prior, we apply joint training to further improve QA performance. Empirical results show that on open-domain question answering datasets like NaturalQuestions and TriviaQA, our method approaches state-of-the-art performance on both generative and extractive inference, all while using much fewer parameters. Furthermore, this strategy allows us to perform hallucination-free inference while conferring significant improvements to the model's ability to rerank relevant passages.

(2 min) The conventional wisdom behind learning deep classification models is to focus on bad-classified examples and ignore well-classified examples that are far from the decision boundary. For instance, when training with cross-entropy loss, examples with higher likelihoods (i.e., well-classified examples) contribute smaller gradients in back-propagation. However, we theoretically show that this common practice hinders representation learning, energy optimization, and the growth of margin. To counteract this deficiency, we propose to reward well-classified examples with additive bonuses to revive their contribution to learning. This counterexample theoretically addresses these three issues. We empirically support this claim by directly verify the theoretical results or through the significant performance improvement with our counterexample on diverse tasks, including image classification, graph classification, and machine translation. Furthermore, this paper shows that because our idea can solve these three issues, we can deal with complex scenarios, such as imbalanced classification, OOD detection, and applications under adversarial attacks.

(2 min) We present state-of-the-art results on morphosyntactic tagging across different varieties of Arabic using fine-tuned pre-trained transformer language models. Our models consistently outperform existing systems in Modern Standard Arabic and all the Arabic dialects we study, achieving 2.6% absolute improvement over the previous state-of-the-art in Modern Standard Arabic, 2.8% in Gulf, 1.6% in Egyptian, and 7.0% in Levantine. We explore different training setups for fine-tuning pre-trained transformer language models, including training data size, the use of external linguistic resources, and the use of annotated data from other dialects in a low-resource scenario. Our results show that strategic fine-tuning using datasets from other high-resource dialects is beneficial for a low-resource dialect. Additionally, we show that high-quality morphological analyzers as external linguistic resources are beneficial especially in low-resource settings.

(2 min) With little to no parallel data available for programming languages, unsupervised methods are well-suited to source code translation. However, the majority of unsupervised machine translation approaches rely on back-translation, a method developed in the context of natural language translation and one that inherently involves training on noisy inputs. Unfortunately, source code is highly sensitive to small changes; a single token can result in compilation failures or erroneous programs, unlike natural languages where small inaccuracies may not change the meaning of a sentence. To address this issue, we propose to leverage an automated unit-testing system to filter out invalid translations, thereby creating a fully tested parallel corpus. We found that fine-tuning an unsupervised model with this filtered data set significantly reduces the noise in the translations so-generated, comfortably outperforming the state-of-the-art for all language pairs studied. In particular, for Java $\to$ Python and Python $\to$ C++ we outperform the best previous methods by more than 16% and 24% respectively, reducing the error rate by more than 35%.

(2 min) Recent developments in machine translation and multilingual text generation have led researchers to adopt trained metrics such as COMET or BLEURT, which treat evaluation as a regression problem and use representations from multilingual pre-trained models such as XLM-RoBERTa or mBERT. Yet studies on related tasks suggest that these models are most efficient when they are large, which is costly and impractical for evaluation. We investigate the trade-off between multilinguality and model capacity with RemBERT, a state-of-the-art multilingual language model, using data from the WMT Metrics Shared Task. We present a series of experiments which show that model size is indeed a bottleneck for cross-lingual transfer, then demonstrate how distillation can help addressing this bottleneck, by leveraging synthetic data generation and transferring knowledge from one teacher to multiple students trained on related languages. Our method yields up to 10.5% improvement over vanilla fine-tuning and reaches 92.6% of RemBERT's performance using only a third of its parameters.

(2 min) Our world is constantly evolving, and so is the content on the web. Consequently, our languages, often said to mirror the world, are dynamic in nature. However, most current contextual language models are static and cannot adapt to changes over time. In this work, we propose a temporal contextual language model called TempoBERT, which uses time as an additional context of texts. Our technique is based on modifying texts with temporal information and performing time masking - specific masking for the supplementary time information. We leverage our approach for the tasks of semantic change detection and sentence time prediction, experimenting on diverse datasets in terms of time, size, genre, and language. Our extensive evaluation shows that both tasks benefit from exploiting time masking.

(2 min) We present a new method LiST for efficient fine-tuning of large pre-trained language models (PLMs) in few-shot learning settings. LiST significantly improves over recent methods that adopt prompt fine-tuning using two key techniques. The first one is the use of self-training to leverage large amounts of unlabeled data for prompt-tuning to significantly boost the model performance in few-shot settings. We use self-training in conjunction with meta-learning for re-weighting noisy pseudo-prompt labels. However, traditional self-training is expensive as it requires updating all the model parameters repetitively. Therefore, we use a second technique for light-weight fine-tuning where we introduce a small number of task-specific adapter parameters that are fine-tuned during self-training while keeping the PLM encoder frozen. This also significantly reduces the overall model footprint across several tasks that can now share a common PLM encoder as backbone for inference. Combining the above techniques, LiST not only improves the model performance for few-shot learning on target domains but also reduces the model memory footprint. We present a comprehensive study on six NLU tasks to validate the effectiveness of LiST. The results show that LiST improves by 35% over classic fine-tuning methods and 6% over prompt-tuning with 96% reduction in number of trainable parameters when fine-tuned with no more than 30 labeled examples from each target domain.

(2 min) Machine learning and neural network models in particular have been improving the state of the art performance on many artificial intelligence related tasks. Neural network models are typically implemented using frameworks that perform gradient based optimization methods to fit a model to a dataset. These frameworks use a technique of calculating derivatives called automatic differentiation (AD) which removes the burden of performing derivative calculations from the model designer. In this report we describe AD, its motivations, and different implementation approaches. We briefly describe dataflow programming as it relates to AD. Lastly, we present example programs that are implemented with Tensorflow and PyTorch, which are two commonly used AD frameworks.

(2 min) While wav2vec 2.0 has been proposed for speech recognition (ASR), it can also be used for speech emotion recognition (SER); its performance can be significantly improved using different fine-tuning strategies. Two baseline methods, vanilla fine-tuning (V-FT) and task adaptive pretraining (TAPT) are first presented. We show that V-FT is able to outperform state-of-the-art models on the IEMOCAP dataset. TAPT, an existing NLP fine-tuning strategy, further improves the performance on SER. We also introduce a novel fine-tuning method termed P-TAPT, which modifies the TAPT objective to learn contextualized emotion representations. Experiments show that P-TAPT performs better than TAPT especially under low-resource settings. Compared to prior works in this literature, our top-line system achieved a 7.4% absolute improvement on unweighted accuracy (UA) over the state-of-the-art performance on IEMOCAP. Our code is publicly available.

(2 min) Building open-domain conversational systems (or chatbots) that produce convincing responses is a recognized challenge. Recent state-of-the-art (SoTA) transformer-based models for the generation of natural language dialogue have demonstrated impressive performance in simulating human-like, single-turn conversations in English. This work investigates, by an empirical study, the potential for transfer learning of such models to Swedish language. DialoGPT, an English language pre-trained model, is adapted by training on three different Swedish language conversational datasets obtained from publicly available sources. Perplexity score (an automated intrinsic language model metric) and surveys by human evaluation were used to assess the performances of the fine-tuned models, with results that indicate that the capacity for transfer learning can be exploited with considerable success. Human evaluators asked to score the simulated dialogue judged over 57% of the chatbot responses to be human-like for the model trained on the largest (Swedish) dataset. We provide the demos and model checkpoints of our English and Swedish chatbots on the HuggingFace platform for public use.

(2 min) Although neural models have shown strong performance in datasets such as SNLI, they lack the ability to generalize out-of-distribution (OOD). In this work, we formulate a few-shot learning setup and examine the effects of natural language explanations on OOD generalization. We leverage the templates in the HANS dataset and construct templated natural language explanations for each template. Although generated explanations show competitive BLEU scores against groundtruth explanations, they fail to improve prediction performance. We further show that generated explanations often hallucinate information and miss key elements that indicate the label.

(2 min) Speech summarization is typically performed by using a cascade of speech recognition and text summarization models. End-to-end modeling of speech summarization models is challenging due to memory and compute constraints arising from long input audio sequences. Recent work in document summarization has inspired methods to reduce the complexity of self-attentions, which enables transformer models to handle long sequences. In this work, we introduce a single model optimized end-to-end for speech summarization. We apply the restricted self-attention technique from text-based models to speech models to address the memory and compute constraints. We demonstrate that the proposed model learns to directly summarize speech for the How-2 corpus of instructional videos. The proposed end-to-end model outperforms the previously proposed cascaded model by 3 points absolute on ROUGE. Further, we consider the spoken language understanding task of predicting concepts from speech inputs and show that the proposed end-to-end model outperforms the cascade model by 4 points absolute F-1.

(2 min) Large scale language models encode rich commonsense knowledge acquired through exposure to massive data during pre-training, but their understanding of entities and their semantic properties is unclear. We probe BERT (Devlin et al., 2019) for the properties of English nouns as expressed by adjectives that do not restrict the reference scope of the noun they modify (as in "red car"), but instead emphasise some inherent aspect ("red strawberry"). We base our study on psycholinguistics datasets that capture the association strength between nouns and their semantic features. We probe BERT using cloze tasks and in a classification setting, and show that the model has marginal knowledge of these features and their prevalence as expressed in these datasets. We discuss factors that make evaluation challenging and impede drawing general conclusions about the models' knowledge of noun properties. Finally, we show that when tested in a fine-tuning setting addressing entailment, BERT successfully leverages the information needed for reasoning about the meaning of adjective-noun constructions outperforming previous methods.

(2 min) In this paper, we present a novel architecture to realize fine-grained style control on the transformer-based text-to-speech synthesis (TransformerTTS). Specifically, we model the speaking style by extracting a time sequence of local style tokens (LST) from the reference speech. The existing content encoder in TransformerTTS is then replaced by our designed cross-attention blocks for fusion and alignment between content and style. As the fusion is performed along with the skip connection, our cross-attention block provides a good inductive bias to gradually infuse the phoneme representation with a given style. Additionally, we prevent the style embedding from encoding linguistic content by randomly truncating LST during training and using wav2vec 2.0 features. Experiments show that with fine-grained style control, our system performs better in terms of naturalness, intelligibility, and style transferability. Our code and samples are publicly available.

(2 min) This paper introduces S3PRL-VC, an open-source voice conversion (VC) framework based on the S3PRL toolkit. In the context of recognition-synthesis VC, self-supervised speech representation (S3R) is valuable in its potential to replace the expensive supervised representation adopted by state-of-the-art VC systems. Moreover, we claim that VC is a good probing task for S3R analysis. In this work, we provide a series of in-depth analyses by benchmarking on the two tasks in VCC2020, namely intra-/cross-lingual any-to-one (A2O) VC, as well as an any-to-any (A2A) setting. We also provide comparisons between not only different S3Rs but also top systems in VCC2020 with supervised representations. Systematic objective and subjective evaluation were conducted, and we show that S3R is comparable with VCC2020 top systems in the A2O setting in terms of similarity, and achieves state-of-the-art in S3R-based A2A VC. We believe the extensive analysis, as well as the toolkit itself, contribute to not only the S3R community but also the VC community. The codebase is now open-sourced.

(2 min) Recent years have witnessed the dramatic growth of paper volumes with plenty of new research papers published every day, especially in the area of computer science. How to glean papers worth reading from the massive literature to do a quick survey or keep up with the latest advancement about a specific research topic has become a challenging task. Existing academic search engines such as Google Scholar return relevant papers by individually calculating the relevance between each paper and query. However, such systems usually omit the prerequisite chains of a research topic and cannot form a meaningful reading path. In this paper, we introduce a new task named Reading Path Generation (RPG) which aims at automatically producing a path of papers to read for a given query. To serve as a research benchmark, we further propose SurveyBank, a dataset consisting of large quantities of survey papers in the field of computer science as well as their citation relationships. Each survey paper contains key phrases extracted from its title and multi-level reading lists inferred from its references. Furthermore, we propose a graph-optimization-based approach for reading path generation which takes the relationship between papers into account. Extensive evaluations demonstrate that our approach outperforms other baselines. A Real-time Reading Path Generation System (RePaGer) has been also implemented with our designed model. To the best of our knowledge, we are the first to target this important research problem. Our source code of RePaGer system and SurveyBank dataset can be found on here.

(2 min) Many machine learning problems involve regressing variables on a non-Euclidean manifold -- e.g. a discrete probability distribution, or the 6D pose of an object. One way to tackle these problems through gradient-based learning is to use a differentiable function that maps arbitrary inputs of a Euclidean space onto the manifold. In this paper, we establish a set of desirable properties for such mapping, and in particular highlight the importance of pre-images connectivity/convexity. We illustrate these properties with a case study regarding 3D rotations. Through theoretical considerations and methodological experiments on a variety of tasks, we review various differentiable mappings on the 3D rotation space, and conjecture about the importance of their local linearity. We show that a mapping based on Procrustes orthonormalization generally performs best among the mappings considered, but that a rotation vector representation might also be suitable when restricted to small angles.

(2 min) Two-stage methods have dominated Human-Object Interaction (HOI) detection for several years. Recently, one-stage HOI detection methods have become popular. In this paper, we aim to explore the essential pros and cons of two-stage and one-stage methods. With this as the goal, we find that conventional two-stage methods mainly suffer from positioning positive interactive human-object pairs, while one-stage methods are challenging to make an appropriate trade-off on multi-task learning, i.e., object detection, and interaction classification. Therefore, a core problem is how to take the essence and discard the dregs from the conventional two types of methods. To this end, we propose a novel one-stage framework with disentangling human-object detection and interaction classification in a cascade manner. In detail, we first design a human-object pair generator based on a state-of-the-art one-stage HOI detector by removing the interaction classification module or head and then design a relatively isolated interaction classifier to classify each human-object pair. Two cascade decoders in our proposed framework can focus on one specific task, detection or interaction classification. In terms of the specific implementation, we adopt a transformer-based HOI detector as our base model. The newly introduced disentangling paradigm outperforms existing methods by a large margin, with a significant relative mAP gain of 9.32% on HICO-Det. The source codes are available at https://github.com/YueLiao/CDN.

(0 min) Our hands reveal important information such as the pulsing of our veins which help us determine the blood pressure, tremors indicative of motor control, or neurodegenerative disorders such as Essential Tremor or Parkinson's disease. The Computerized Classification of Micro-Motions in the hand using waveforms from mobile phone videos is a novel method that uses Eulerian Video Magnification, Skeletonization, Heatmapping, and the kNN machine learning model to detect the micro-motions in the human hand, synthesize their waveforms, and classify these. The pre-processing is achieved by using Eulerian Video Magnification, Skeletonization, and Heat-mapping to magnify the micro-motions, landmark essential features of the hand, and determine the extent of motion, respectively. Following pre-processing, the visible motions are manually labeled by appropriately grouping pixels to represent a particular label correctly. These labeled motions of the pixels are converted into waveforms. Finally, these waveforms are classified into four categories - hand or finger movements, vein movement, background motion, and movement of the rest of the body due to respiration using the kNN model. The final accuracy obtained was around 92 percent.

(2 min) In this short report, we present a simple, yet effective approach to editing real images via generative adversarial networks (GAN). Unlike previous techniques, that treat all editing tasks as an operation that affects pixel values in the entire image in our approach we cut up the image into a set of smaller segments. For those segments corresponding latent codes of a generative network can be estimated with greater accuracy due to the lower number of constraints. When codes are altered by the user the content in the image is manipulated locally while the rest of it remains unaffected. Thanks to this property the final edited image better retains the original structures and thus helps to preserve natural look.

(2 min) Randomized Smoothing (RS) is a promising method for obtaining robustness certificates by evaluating a base model under noise. In this work, we: (i) theoretically motivate why ensembles are a particularly suitable choice as base models for RS, and (ii) empirically confirm this choice, obtaining state-of-the-art results in multiple settings. The key insight of our work is that the reduced variance of ensembles over the perturbations introduced in RS leads to significantly more consistent classifications for a given input. This, in turn, leads to substantially increased certifiable radii for samples close to the decision boundary. Additionally, we introduce key optimizations which enable an up to 55-fold decrease in sample complexity of RS, thus drastically reducing its computational overhead. Experimentally, we show that ensembles of only 3 to 10 classifiers consistently improve on their strongest constituting model with respect to their average certified radius (ACR) by 5% to 21% on both CIFAR10 and ImageNet, achieving a new state-of-the-art ACR of 0.86 and 1.11, respectively. We release all code and models required to reproduce our results upon publication.

(2 min) Together with critical editions and translations, commentaries are one of the main genres of publication in literary and textual scholarship, and have a century-long tradition. Yet, the exploitation of thousands of digitized historical commentaries was hitherto hindered by the poor quality of Optical Character Recognition (OCR), especially on commentaries to Greek texts. In this paper, we evaluate the performances of two pipelines suitable for the OCR of historical classical commentaries. Our results show that Kraken + Ciaconna reaches a substantially lower character error rate (CER) than Tesseract/OCR-D on commentary sections with high density of polytonic Greek text (average CER 7% vs. 13%), while Tesseract/OCR-D is slightly more accurate than Kraken + Ciaconna on text sections written predominantly in Latin script (average CER 8.2% vs. 8.4%). As part of this paper, we also release GT4HistComment, a small dataset with OCR ground truth for 19th classical commentaries and Pogretra, a large collection of training data and pre-trained models for a wide variety of ancient Greek typefaces.

(2 min) The state-of-the-art deep neural networks are vulnerable to common corruptions (e.g., input data degradations, distortions, and disturbances caused by weather changes, system error, and processing). While much progress has been made in analyzing and improving the robustness of models in image understanding, the robustness in video understanding is largely unexplored. In this paper, we establish a corruption robustness benchmark, Mini Kinetics-C and Mini SSV2-C, which considers temporal corruptions beyond spatial corruptions in images. We make the first attempt to conduct an exhaustive study on the corruption robustness of established CNN-based and Transformer-based spatial-temporal models. The study provides some guidance on robust model design and training: Transformer-based model performs better than CNN-based models on corruption robustness; the generalization ability of spatial-temporal models implies robustness against temporal corruptions; model corruption robustness (especially robustness in the temporal domain) enhances with computational cost and model capacity, which may contradict the current trend of improving the computational efficiency of models. Moreover, we find the robustness intervention for image-related tasks (e.g., training models with noise) may not work for spatial-temporal models.

(2 min) 3D object detection often involves complicated training and testing pipelines, which require substantial domain knowledge about individual datasets. Inspired by recent non-maximum suppression-free 2D object detection models, we propose a 3D object detection architecture on point clouds. Our method models 3D object detection as message passing on a dynamic graph, generalizing the DGCNN framework to predict a set of objects. In our construction, we remove the necessity of post-processing via object confidence aggregation or non-maximum suppression. To facilitate object detection from sparse point clouds, we also propose a set-to-set distillation approach customized to 3D detection. This approach aligns the outputs of the teacher model and the student model in a permutation-invariant fashion, significantly simplifying knowledge distillation for the 3D detection task. Our method achieves state-of-the-art performance on autonomous driving benchmarks. We also provide abundant analysis of the detection model and distillation framework.

(2 min) Neural Radiance Fields (NeRF) have recently demonstrated photo-realistic results for the task of novel view synthesis. In this paper, we propose to apply novel view synthesis to the robot relocalization problem: we demonstrate improvement of camera pose regression thanks to an additional synthetic dataset rendered by the NeRF class of algorithm. To avoid spawning novel views in irrelevant places we selected virtual camera locations from NeRF internal representation of the 3D geometry of the scene. We further improved localization accuracy of pose regressors using synthesized realistic and geometry consistent images as data augmentation during training. At the time of publication, our approach improved state of the art with a 60% lower error on Cambridge Landmarks and 7-scenes datasets. Hence, the resulting accuracy becomes comparable to structure-based methods, without any architecture modification or domain adaptation constraints. Since our method allows almost infinite generation of training data, we investigated limitations of camera pose regression depending on size and distribution of data used for training on public benchmarks. We concluded that pose regression accuracy is mostly bounded by relatively small and biased datasets rather than capacity of the pose regression model to solve the localization task.

(2 min) Real-world machine learning systems need to analyze test data that may differ from training data. In K-way classification, this is crisply formulated as open-set recognition, core to which is the ability to discriminate open-set data outside the K closed-set classes. Two conceptually elegant ideas for open-set discrimination are: 1) discriminatively learning an open-vs-closed binary discriminator by exploiting some outlier data as the open-set, and 2) unsupervised learning the closed-set data distribution with a GAN, using its discriminator as the open-set likelihood function. However, the former generalizes poorly to diverse open test data due to overfitting to the training outliers, which are unlikely to exhaustively span the open-world. The latter does not work well, presumably due to the instable training of GANs. Motivated by the above, we propose OpenGAN, which addresses the limitation of each approach by combining them with several technical insights. First, we show that a carefully selected GAN-discriminator on some real outlier data already achieves the state-of-the-art. Second, we augment the available set of real open training examples with adversarially synthesized "fake" data. Third and most importantly, we build the discriminator over the features computed by the closed-world K-way networks. This allows OpenGAN to be implemented via a lightweight discriminator head built on top of an existing K-way network. Extensive experiments show that OpenGAN significantly outperforms prior open-set methods.

(2 min) Domain generalization (DG) aims to learn from multiple known source domains a model that can generalize well to unknown target domains. The existing DG methods usually rely on shared multi-source data fusion for generalizable model training. However, tremendous data is distributed across lots of places nowadays that can not be shared due to privacy policies, especially in some crucial areas like finance and medical care. A dilemma is thus raised between real-world data privacy protection and simultaneous multi-source semantic learning with the shared data. In this paper, we investigate a separated domain generalization task with separated source datasets that can only be used locally, which is vital for real-world privacy protection. We propose a novel solution called Collaborative Semantic Aggregation and Calibration (CSAC) to enable this challenging task. To fully absorb multi-source semantic information while avoiding unsafe data fusion, we first conduct data-free semantic aggregation by fusing the models trained on the separated domains layer-by-layer. To address semantic dislocation caused by domain shift, we further design cross-layer semantic calibration with an attention mechanism to align each semantic level and enhance domain invariance. We unify multi-source semantic learning and alignment in a collaborative way by repeating the semantic aggregation and calibration alternately, keeping each dataset localized, and privacy is thus carefully protected. Extensive experiments show the significant performance of our method in addressing this challenging task, which is even comparable to the previous DG methods with shared data.

(2 min) In the recent years, many methods demonstrated the ability of neural networks to learn depth and pose changes in a sequence of images, using only self-supervision as the training signal. Whilst the networks achieve good performance, the often over-looked detail is that due to the inherent ambiguity of monocular vision they predict depth up to an unknown scaling factor. The scaling factor is then typically obtained from the LiDAR ground truth at test time, which severely limits practical applications of these methods. In this paper, we show that incorporating prior information about the camera configuration and the environment, we can remove the scale ambiguity and predict depth directly, still using the self-supervised formulation and not relying on any additional sensors.

(2 min) Accurately maintaining digital street maps is labor-intensive. To address this challenge, much work has studied automatically processing geospatial data sources such as GPS trajectories and satellite images to reduce the cost of maintaining digital maps. An end-to-end map update system would first process geospatial data sources to extract insights, and second leverage those insights to update and improve the map. However, prior work largely focuses on the first step of this pipeline: these map extraction methods infer road networks from scratch given geospatial data sources (in effect creating entirely new maps), but do not address the second step of leveraging this extracted information to update the existing digital map data. In this paper, we first explain why current map extraction techniques yield low accuracy when extended to update existing maps. We then propose a novel method that leverages the progression of satellite imagery over time to substantially improve accuracy. Our approach first compares satellite images captured at different times to identify portions of the physical road network that have visibly changed, and then updates the existing map accordingly. We show that our change-based approach reduces map update error rates four-fold.

(2 min) We present a novel point-based, differentiable neural rendering pipeline for scene refinement and novel view synthesis. The input are an initial estimate of the point cloud and the camera parameters. The output are synthesized images from arbitrary camera poses. The point cloud rendering is performed by a differentiable renderer using multi-resolution one-pixel point rasterization. Spatial gradients of the discrete rasterization are approximated by the novel concept of ghost geometry. After rendering, the neural image pyramid is passed through a deep neural network for shading calculations and hole-filling. A differentiable, physically-based tonemapper then converts the intermediate output to the target image. Since all stages of the pipeline are differentiable, we optimize all of the scene's parameters i.e. camera model, camera pose, point position, point color, environment map, rendering network weights, vignetting, camera response function, per image exposure, and per image white balance. We show that our system is able to synthesize sharper and more consistent novel views than existing approaches because the initial reconstruction is refined during training. The efficient one-pixel point rasterization allows us to use arbitrary camera models and display scenes with well over 100M points in real time.

(2 min) The current state-of-the-art in user mobility research has extensively relied on open-source mobility traces captured from pedestrian and vehicular activity through a variety of communication technologies as users engage in a wide-range of applications, including connected healthcare, localization, social media, e-commerce, etc. Most of these traces are feature-rich and diverse, not only in the information they provide, but also in how they can be used and leveraged. This diversity poses two main challenges for researchers and practitioners who wish to make use of available mobility datasets. First, it is quite difficult to get a bird's eye view of the available traces without spending considerable time looking them up. Second, once they have found the traces, they still need to figure out whether the traces are adequate to their needs. The purpose of this survey is three-fold. It proposes a taxonomy to classify open-source mobility traces including their mobility mode, data source and collection technology. It then uses the proposed taxonomy to classify existing open-source mobility traces and finally, highlights three case studies using popular publicly available datasets to showcase how our taxonomy can tease out feature sets in traces to help determine their applicability to specific use-cases.

(2 min) Multi-modal domain translation typically refers to synthesizing a novel image that inherits certain localized attributes from a 'content' image (e.g. layout, semantics, or geometry), and inherits everything else (e.g. texture, lighting, sometimes even semantics) from a 'style' image. The dominant approach to this task is attempting to learn disentangled 'content' and 'style' representations from scratch. However, this is not only challenging, but ill-posed, as what users wish to preserve during translation varies depending on their goals. Motivated by this inherent ambiguity, we define 'content' based on conditioning information extracted by off-the-shelf pre-trained models. We then train our style extractor and image decoder with an easy to optimize set of reconstruction objectives. The wide variety of high-quality pre-trained models available and simple training procedure makes our approach straightforward to apply across numerous domains and definitions of 'content'. Additionally it offers intuitive control over which aspects of 'content' are preserved across domains. We evaluate our method on traditional, well-aligned, datasets such as CelebA-HQ, and propose two novel datasets for evaluation on more complex scenes: ClassicTV and FFHQ-Wild. Our approach, Sensorium, enables higher quality domain translation for more complex scenes.

(2 min) This paper tackles the problem of parts-aware point cloud generation. Unlike existing works which require the point cloud to be segmented into parts a priori, our parts-aware editing and generation is performed in an unsupervised manner. We achieve this with a simple modification of the Variational Auto-Encoder which yields a joint model of the point cloud itself along with a schematic representation of it as a combination of shape primitives. In particular, we introduce a latent representation of the point cloud which can be decomposed into a disentangled representation for each part of the shape. These parts are in turn disentangled into both a shape primitive and a point cloud representation, along with a standardising transformation to a canonical coordinate system. The dependencies between our standardising transformations preserve the spatial dependencies between the parts in a manner which allows meaningful parts-aware point cloud generation and shape editing. In addition to the flexibility afforded by our disentangled representation, the inductive bias introduced by our joint modelling approach yields the state-of-the-art experimental results on the ShapeNet dataset.

(0 min) Few-shot semantic segmentation models aim to segment images after learning from only a few annotated examples. A key challenge for them is how to avoid overfitting because limited training data is available. While prior works usually limited the overall model capacity to alleviate overfitting, this hampers segmentation accuracy. We demonstrate how to increase overall model capacity to achieve improved performance, by introducing objectness, which is class-agnostic and so not prone to overfitting, for complementary use with class-specific features. Extensive experiments demonstrate the versatility of our simple approach of introducing objectness for different base architectures that rely on different data loaders and training schedules (DENet, PFENet) as well as with different backbone models (ResNet-50, ResNet-101 and HRNetV2-W48). Given only one annotated example of an unseen category, experiments show that our method outperforms state-of-art methods with respect to mIoU by at least 4.7% and 1.5% on PASCAL-5i and COCO-20i respectively.

(0 min) We propose a segmentation-based bounding box generation method for omnidirectional pedestrian detection that enables detectors to tightly fit bounding boxes to pedestrians without omnidirectional images for training. Due to the wide angle of view, omnidirectional cameras are more cost-effective than standard cameras and hence suitable for large-scale monitoring. The problem of using omnidirectional cameras for pedestrian detection is that the performance of standard pedestrian detectors is likely to be substantially degraded because pedestrians' appearance in omnidirectional images may be rotated to any angle. Existing methods mitigate this issue by transforming images during inference. However, the transformation substantially degrades the detection accuracy and speed. A recently proposed method obviates the transformation by training detectors with omnidirectional images, which instead incurs huge annotation costs. To obviate both the transformation and annotation works, we leverage an existing large-scale object detection dataset. We train a detector with rotated images and tightly fitted bounding box annotations generated from the segmentation annotations in the dataset, resulting in detecting pedestrians in omnidirectional images with tightly fitted bounding boxes. We also develop pseudo-fisheye distortion augmentation, which further enhances the performance. Extensive analysis shows that our detector successfully fits bounding boxes to pedestrians and demonstrates substantial performance improvement.

(0 min) We present ARCH++, an image-based method to reconstruct 3D avatars with arbitrary clothing styles. Our reconstructed avatars are animation-ready and highly realistic, in both the visible regions from input views and the unseen regions. While prior work shows great promise of reconstructing animatable clothed humans with various topologies, we observe that there exist fundamental limitations resulting in sub-optimal reconstruction quality. In this paper, we revisit the major steps of image-based avatar reconstruction and address the limitations with ARCH++. First, we introduce an end-to-end point based geometry encoder to better describe the semantics of the underlying 3D human body, in replacement of previous hand-crafted features. Second, in order to address the occupancy ambiguity caused by topological changes of clothed humans in the canonical pose, we propose a co-supervising framework with cross-space consistency to jointly estimate the occupancy in both the posed and canonical spaces. Last, we use image-to-image translation networks to further refine detailed geometry and texture on the reconstructed surface, which improves the fidelity and consistency across arbitrary viewpoints. In the experiments, we demonstrate improvements over the state of the art on both public benchmarks and user studies in reconstruction quality and realism.

(0 min) This report describes our solution to the VALUE Challenge 2021 in the captioning task. Our solution, named CLIP4Caption++, is built on X-Linear/X-Transformer, which is an advanced model with encoder-decoder architecture. We make the following improvements on the proposed CLIP4Caption++: We employ an advanced encoder-decoder model architecture X-Transformer as our main framework and make the following improvements: 1) we utilize three strong pre-trained CLIP models to extract the text-related appearance visual features. 2) we adopt the TSN sampling strategy for data enhancement. 3) we involve the video subtitle information to provide richer semantic information. 3) we introduce the subtitle information, which fuses with the visual features as guidance. 4) we design word-level and sentence-level ensemble strategies. Our proposed method achieves 86.5, 148.4, 64.5 CIDEr scores on VATEX, YC2C, and TVC datasets, respectively, which shows the superior performance of our proposed CLIP4Caption++ on all three datasets.

(0 min) Recently, Zhang et al. (2021) developed a new neural network architecture based on $\ell_\infty$-distance functions, which naturally possesses certified robustness by its construction. Despite the excellent theoretical properties, the model so far can only achieve comparable performance to conventional networks. In this paper, we significantly boost the certified robustness of $\ell_\infty$-distance nets through a careful analysis of its training process. In particular, we show the $\ell_p$-relaxation, a crucial way to overcome the non-smoothness of the model, leads to an unexpected large Lipschitz constant at the early training stage. This makes the optimization insufficient using hinge loss and produces sub-optimal solutions. Given these findings, we propose a simple approach to address the issues above by using a novel objective function that combines a scaled cross-entropy loss with clipped hinge loss. Our experiments show that using the proposed training strategy, the certified accuracy of $\ell_\infty$-distance net can be dramatically improved from 33.30% to 40.06% on CIFAR-10 ($\epsilon=8/255$), meanwhile significantly outperforming other approaches in this area. Such a result clearly demonstrates the effectiveness and potential of $\ell_\infty$-distance net for certified robustness.

(0 min) Multi-object tracking (MOT) aims at estimating bounding boxes and identities of objects in videos. Most methods obtain identities by associating detection boxes whose scores are higher than a threshold. The objects with low detection scores, e.g. occluded objects, are simply thrown away, which brings non-negligible true object missing and fragmented trajectories. To solve this problem, we present a simple, effective and generic association method, called BYTE, tracking BY associaTing Every detection box instead of only the high score ones. For the low score detection boxes, we utilize their similarities with tracklets to recover true objects and filter out the background detections. We apply BYTE to 9 different state-of-the-art trackers and achieve consistent improvement on IDF1 score ranging from 1 to 10 points. To put forwards the state-of-the-art performance of MOT, we design a simple and strong tracker, named ByteTrack. For the first time, we achieve 80.3 MOTA, 77.3 IDF1 and 63.1 HOTA on the test set of MOT17 with 30 FPS running speed on a single V100 GPU. The source code, pre-trained models with deploy versions and tutorials of applying to other trackers are released at \url{https://github.com/ifzhang/ByteTrack}.

(0 min) In the domain of Visual Question Answering (VQA), studies have shown improvement in users' mental model of the VQA system when they are exposed to examples of how these systems answer certain Image-Question (IQ) pairs. In this work, we show that showing controlled counterfactual image-question examples are more effective at improving the mental model of users as compared to simply showing random examples. We compare a generative approach and a retrieval-based approach to show counterfactual examples. We use recent advances in generative adversarial networks (GANs) to generate counterfactual images by deleting and inpainting certain regions of interest in the image. We then expose users to changes in the VQA system's answer on those altered images. To select the region of interest for inpainting, we experiment with using both human-annotated attention maps and a fully automatic method that uses the VQA system's attention values. Finally, we test the user's mental model by asking them to predict the model's performance on a test counterfactual image. We note an overall improvement in users' accuracy to predict answer change when shown counterfactual explanations. While realistic retrieved counterfactuals obviously are the most effective at improving the mental model, we show that a generative approach can also be equally effective.

(0 min) Estimating the motion of the camera together with the 3D structure of the scene from a monocular vision system is a complex task that often relies on the so-called scene rigidity assumption. When observing a dynamic environment, this assumption is violated which leads to an ambiguity between the ego-motion of the camera and the motion of the objects. To solve this problem, we present a self-supervised learning framework for 3D object motion field estimation from monocular videos. Our contributions are two-fold. First, we propose a two-stage projection pipeline to explicitly disentangle the camera ego-motion and the object motions with dynamics attention module, called DAM. Specifically, we design an integrated motion model that estimates the motion of the camera and object in the first and second warping stages, respectively, controlled by the attention module through a shared motion encoder. Second, we propose an object motion field estimation through contrastive sample consensus, called CSAC, taking advantage of weak semantic prior (bounding box from an object detector) and geometric constraints (each object respects the rigid body motion model). Experiments on KITTI, Cityscapes, and Waymo Open Dataset demonstrate the relevance of our approach and show that our method outperforms state-of-the-art algorithms for the tasks of self-supervised monocular depth estimation, object motion segmentation, monocular scene flow estimation, and visual odometry.

(0 min) Evidence from cognitive psychology suggests that understanding spatio-temporal object interactions and dynamics can be essential for recognizing actions in complex videos. Therefore, action recognition models are expected to benefit from explicit modeling of objects, including their appearance, interaction, and dynamics. Recently, video transformers have shown great success in video understanding, exceeding CNN performance. Yet, existing video transformer models do not explicitly model objects. In this work, we present Object-Region Video Transformers (ORViT), an \emph{object-centric} approach that extends video transformer layers with a block that directly incorporates object representations. The key idea is to fuse object-centric spatio-temporal representations throughout multiple transformer layers. Our ORViT block consists of two object-level streams: appearance and dynamics. In the appearance stream, an ``Object-Region Attention'' element applies self-attention over the patches and \emph{object regions}. In this way, visual object regions interact with uniform patch tokens and enrich them with contextualized object information. We further model object dynamics via a separate ``Object-Dynamics Module'', which captures trajectory interactions, and show how to integrate the two streams. We evaluate our model on standard and compositional action recognition on Something-Something V2, standard action recognition on Epic-Kitchen100 and Diving48, and spatio-temporal action detection on AVA. We show strong improvement in performance across all tasks and datasets considered, demonstrating the value of a model that incorporates object representations into a transformer architecture. For code and pretrained models, visit the project page at https://roeiherz.github.io/ORViT/.

(0 min) Few-shot learning aims to recognize novel classes from a few examples. Although significant progress has been made in the image domain, few-shot video classification is relatively unexplored. We argue that previous methods underestimate the importance of video feature learning and propose to learn spatiotemporal features using a 3D CNN. Proposing a two-stage approach that learns video features on base classes followed by fine-tuning the classifiers on novel classes, we show that this simple baseline approach outperforms prior few-shot video classification methods by over 20 points on existing benchmarks. To circumvent the need of labeled examples, we present two novel approaches that yield further improvement. First, we leverage tag-labeled videos from a large dataset using tag retrieval followed by selecting the best clips with visual similarities. Second, we learn generative adversarial networks that generate video features of novel classes from their semantic embeddings. Moreover, we find existing benchmarks are limited because they only focus on 5 novel classes in each testing episode and introduce more realistic benchmarks by involving more novel classes, i.e. few-shot learning, as well as a mixture of novel and base classes, i.e. generalized few-shot learning. The experimental results show that our retrieval and feature generation approach significantly outperform the baseline approach on the new benchmarks.

(0 min) Recent work has argued that classification losses utilizing softmax cross-entropy are superior not only for fixed-set classification tasks, but also by outperforming losses developed specifically for open-set tasks including few-shot learning and retrieval. Softmax classifiers have been studied using different embedding geometries -- Euclidean, hyperbolic, and spherical -- and claims have been made about the superiority of one or another, but they have not been systematically compared with careful controls. We conduct an empirical investigation of embedding geometry on softmax losses for a variety of fixed-set classification and image retrieval tasks. An interesting property observed for the spherical losses lead us to propose a probabilistic classifier based on the von Mises-Fisher distribution, and we show that it is competitive with state-of-the-art methods while producing improved out-of-the-box calibration. We provide guidance regarding the trade-offs between losses and how to choose among them.

(0 min) Neural Architecture Search (NAS) has been pivotal in finding optimal network configurations for Convolution Neural Networks (CNNs). While many methods explore NAS from a global search-space perspective, the employed optimization schemes typically require heavy computational resources. This work introduces a method that is efficient in computationally constrained environments by examining the micro-search space of channel size. In tackling channel-size optimization, we design an automated algorithm to extract the dependencies within different connected layers of the network. In addition, we introduce the idea of knowledge distillation, which enables preservation of trained weights, admist trials where the channel sizes are changing. Further, since the standard performance indicators (accuracy, loss) fail to capture the performance of individual network components (providing an overall network evaluation), we introduce a novel metric that highly correlates with test accuracy and enables analysis of individual network layers. Combining dependency extraction, metrics, and knowledge distillation, we introduce an efficient searching algorithm, with simulated annealing inspired stochasticity, and demonstrate its effectiveness in finding optimal architectures that outperform baselines by a large margin.

(0 min) Semi-Supervised Learning (SSL) can decrease the required amount of labeled image data and thus the cost for deep learning. Most SSL methods assume a clear distinction between classes, but class boundaries are often ambiguous in real-world datasets due to intra- or interobserver variability. This ambiguity of annotations must be addressed as it will otherwise limit the performance of SSL and deep learning in general due to inconsistent label information. We propose Semi-Supervised Classification & Clustering (S2C2) which can extend many deep SSL algorithms. S2C2 automatically estimates the ambiguity of an image and applies the respective SSL algorithm as a classification to certainly labeled data while partitioning the ambiguous data into clusters of visual similar images. We show that S2C2 results in a 7.6% better F1-score for classifications and 7.9% lower inner distance of clusters on average across multiple SSL algorithms and datasets. Moreover, the output of S2C2 can be used to decrease the ambiguity of labels with the help of human experts. Overall, a combination of Semi-Supervised Learning with our method S2C2 leads to better handling of ambiguous labels and thus real-world datasets.

(0 min) Robots in our daily surroundings are increasing day by day. Their usability and acceptability largely depend on their explicit and implicit interaction capability with fellow human beings. As a result, social behavior is one of the most sought-after qualities that a robot can possess. However, there is no specific aspect and/or feature that defines socially acceptable behavior and it largely depends on the situation, application, and society. In this article, we investigate one such social behavior for collocated robots. Imagine a group of people is interacting with each other and we want to join the group. We as human beings do it in a socially acceptable manner, i.e., within the group, we do position ourselves in such a way that we can participate in the group activity without disturbing/obstructing anybody. To possess such a quality, first, a robot needs to determine the formation of the group and then determine a position for itself, which we humans do implicitly. The theory of f-formation can be utilized for this purpose. As the types of formations can be very diverse, detecting the social groups is not a trivial task. In this article, we provide a comprehensive survey of the existing work on social interaction and group detection using f-formation for robotics and other applications. We also put forward a novel holistic survey framework combining all the possible concerns and modules relevant to this problem. We define taxonomies based on methods, camera views, datasets, detection capabilities and scale, evaluation approaches, and application areas. We discuss certain open challenges and limitations in current literature along with possible future research directions based on this framework. In particular, we discuss the existing methods/techniques and their relative merits and demerits, applications, and provide a set of unsolved but relevant problems in this domain.

(0 min) Limited availability of large image datasets is a major issue in the development of accurate and generalizable machine learning methods in medicine. The limitations in the amount of data are mainly due to the use of different acquisition protocols, different hardware, and data privacy. At the same time, training a classification model on a small dataset leads to a poor generalization quality of the model. To overcome this issue, a combination of various image datasets of different provenance is often used, e.g., multi-site studies. However, if an additional dataset does not include all classes of the task, the learning of the classification model can be biased to the device or place of acquisition. This is especially the case for Magnetic Resonance (MR) images, where different MR scanners introduce a bias that limits the performance of the model. In this paper, we present a novel method that learns to ignore the scanner-related features present in the images, while learning features relevant for the classification task. We focus on a real-world scenario, where only a small dataset provides images of all classes. We exploit this circumstance by introducing specific additional constraints on the latent space, which lead the focus on disease-related rather than scanner-specific features. Our method Learn to Ignore outperforms state-of-the-art domain adaptation methods on a multi-site MRI dataset on a classification task between Multiple Sclerosis patients and healthy subjects.

(0 min) CNN (Convolutional Neural Network) models have been successfully used for segmentation of the left ventricle (LV) in cardiac MRI (Magnetic Resonance Imaging), providing clinical measurements. In practice, two questions arise with deployment of CNNs: 1) when is it better to use a shallow model instead of a deeper one? 2) how the size of a dataset might change the network performance? We propose a framework to answer them, by experimenting with deep and shallow versions of three U-Net families, trained from scratch in six subsets varying from 100 to 10,000 images, different network sizes, learning rates and regularization values. 1620 models were evaluated using 5-fold cross-validation by loss and DICE. The results indicate that: sample size affects performance more than architecture or hyper-parameters; in small samples the performance is more sensitive to hyper-parameters than architecture; the performance difference between shallow and deeper networks is not the same across families.

(0 min) Automatic font generation based on deep learning has aroused a lot of interest in the last decade. However, only a few recently-reported approaches are capable of directly generating vector glyphs and their results are still far from satisfactory. In this paper, we propose a novel method, DeepVecFont, to effectively resolve this problem. Using our method, for the first time, visually-pleasing vector glyphs whose quality and compactness are both comparable to human-designed ones can be automatically generated. The key idea of our DeepVecFont is to adopt the techniques of image synthesis, sequence modeling and differentiable rasterization to exhaustively exploit the dual-modality information (i.e., raster images and vector outlines) of vector fonts. The highlights of this paper are threefold. First, we design a dual-modality learning strategy which utilizes both image-aspect and sequence-aspect features of fonts to synthesize vector glyphs. Second, we provide a new generative paradigm to handle unstructured data (e.g., vector glyphs) by randomly sampling plausible synthesis results to get the optimal one which is further refined under the guidance of generated structured data (e.g., glyph images). Finally, qualitative and quantitative experiments conducted on a publicly-available dataset demonstrate that our method obtains high-quality synthesis results in the applications of vector font generation and interpolation, significantly outperforming the state of the art.

(0 min) The adversarial attack literature contains a myriad of algorithms for crafting perturbations which yield pathological behavior in neural networks. In many cases, multiple algorithms target the same tasks and even enforce the same constraints. In this work, we show that different attack algorithms produce adversarial examples which are distinct not only in their effectiveness but also in how they qualitatively affect their victims. We begin by demonstrating that one can determine the attack algorithm that crafted an adversarial example. Then, we leverage recent advances in parameter-space saliency maps to show, both visually and quantitatively, that adversarial attack algorithms differ in which parts of the network and image they target. Our findings suggest that prospective adversarial attacks should be compared not only via their success rates at fooling models but also via deeper downstream effects they have on victims.

(0 min) Machine learning image classifiers are susceptible to adversarial and corruption perturbations. Adding imperceptible noise to images can lead to severe misclassifications of the machine learning model. Using $L_p$-norms for measuring the size of the noise fails to capture human similarity perception, which is why optimal transport based distance measures like the Wasserstein metric are increasingly being used in the field of adversarial robustness. Verifying the robustness of classifiers using the Wasserstein metric can be achieved by proving the absence of adversarial examples (certification) or proving their presence (attack). In this work we present a framework based on the work by Levine and Feizi, which allows us to transfer existing certification methods for convex polytopes or $L_1$-balls to the Wasserstein threat model. The resulting certification can be complete or incomplete, depending on whether convex polytopes or $L_1$-balls were chosen. Additionally, we present a new Wasserstein adversarial attack that is projected gradient descent based and which has a significantly reduced computational burden compared to existing attack approaches.

(0 min) We investigate the ability to detect slag formations in images from inside a Grate-Kiln system furnace with two deep convolutional neural networks. The conditions inside the furnace cause occasional obstructions of the camera view. Our approach suggests dealing with this problem by introducing a convLSTM-layer in the deep convolutional neural network. The results show that it is possible to achieve sufficient performance to automate the decision of timely countermeasures in the industrial operational setting. Furthermore, the addition of the convLSTM-layer results in fewer outlying predictions and a lower running variance of the fraction of detected slag in the image time series.

(0 min) Classification of images within the compressed domain offers significant benefits. These benefits include reduced memory and computational requirements of a classification system. This paper proposes two such methods as a proof of concept: The first classifies within the JPEG image transform domain (i.e. DCT transform data); the second classifies the JPEG compressed binary bitstream directly. These two methods are implemented using Residual Network CNNs and an adapted Vision Transformer. Top-1 accuracy of approximately 70% and 60% were achieved using these methods respectively when classifying the Caltech C101 database. Although these results are significantly behind the state of the art for classification for this database (~95%), it illustrates the first time direct bitstream image classification has been achieved. This work confirms that direct bitstream image classification is possible and could be utilised in a first pass database screening of a raw bitstream (within a wired or wireless network) or where computational, memory and bandwidth requirements are severely restricted.

(0 min) Hyperspectral imaging has been increasingly used for underwater survey applications over the past years. As many hyperspectral cameras work as push-broom scanners, their use is usually limited to the creation of photo-mosaics based on a flat surface approximation and by interpolating the camera pose from dead-reckoning navigation. Yet, because of drift in the navigation and the mostly wrong flat surface assumption, the quality of the obtained photo-mosaics is often too low to support adequate analysis.In this paper we present an initial method for creating hyperspectral 3D reconstructions of underwater environments. By fusing the data gathered by a classical RGB camera, an inertial navigation system and a hyperspectral push-broom camera, we show that the proposed method creates highly accurate 3D reconstructions with hyperspectral textures. We propose to combine techniques from simultaneous localization and mapping, structure-from-motion and 3D reconstruction and advantageously use them to create 3D models with hyperspectral texture, allowing us to overcome the flat surface assumption and the classical limitation of dead-reckoning navigation.

(0 min) Deep learning has been successfully applied to many classification problems including underwater challenges. However, a long-standing issue with deep learning is the need for large and consistently labeled datasets. Although current approaches in semi-supervised learning can decrease the required amount of annotated data by a factor of 10 or even more, this line of research still uses distinct classes. For underwater classification, and uncurated real-world datasets in general, clean class boundaries can often not be given due to a limited information content in the images and transitional stages of the depicted objects. This leads to different experts having different opinions and thus producing fuzzy labels which could also be considered ambiguous or divergent. We propose a novel framework for handling semi-supervised classifications of such fuzzy labels. It is based on the idea of overclustering to detect substructures in these fuzzy labels. We propose a novel loss to improve the overclustering capability of our framework and show the benefit of overclustering for fuzzy labels. We show that our framework is superior to previous state-of-the-art semi-supervised methods when applied to real-world plankton data with fuzzy labels. Moreover, we acquire 5 to 10\% more consistent predictions of substructures.

(0 min) Though a number of point cloud learning methods have been proposed to handle unordered points, most of them are supervised and require labels for training. By contrast, unsupervised learning of point cloud data has received much less attention to date. In this paper, we propose a simple yet effective approach for unsupervised point cloud learning. In particular, we identify a very useful transformation which generates a good contrastive version of an original point cloud. They make up a pair. After going through a shared encoder and a shared head network, the consistency between the output representations are maximized with introducing two variants of contrastive losses to respectively facilitate downstream classification and segmentation. To demonstrate the efficacy of our method, we conduct experiments on three downstream tasks which are 3D object classification (on ModelNet40 and ModelNet10), shape part segmentation (on ShapeNet Part dataset) as well as scene segmentation (on S3DIS). Comprehensive results show that our unsupervised contrastive representation learning enables impressive outcomes in object classification and semantic segmentation. It generally outperforms current unsupervised methods, and even achieves comparable performance to supervised methods. Our source codes will be made publicly available.

(0 min) In this paper, we consider the task of space-time video super-resolution (ST-VSR), which simultaneously increases the spatial resolution and frame rate for a given video. However, existing methods typically suffer from difficulties in how to efficiently leverage information from a large range of neighboring frames or avoiding the speed degradation in the inference using deformable ConvLSTM strategies for alignment. % Some recent LSTM-based ST-VSR methods have achieved promising results. To solve the above problem of the existing methods, we propose a coarse-to-fine bidirectional recurrent neural network instead of using ConvLSTM to leverage knowledge between adjacent frames. Specifically, we first use bi-directional optical flow to update the hidden state and then employ a Feature Refinement Module (FRM) to refine the result. Since we could fully utilize a large range of neighboring frames, our method leverages local and global information more effectively. In addition, we propose an optical flow-reuse strategy that can reuse the intermediate flow of adjacent frames, which considerably reduces the computation burden of frame alignment compared with existing LSTM-based designs. Extensive experiments demonstrate that our optical-flow-reuse-based bidirectional recurrent network(OFR-BRN) is superior to the state-of-the-art methods both in terms of accuracy and efficiency.

(0 min) Graphic design is ubiquitous in people's daily lives. For graphic design, the most time-consuming task is laying out various components in the interface. Repetitive manual layout design will waste a lot of time for professional graphic designers. Existing templates are usually rudimentary and not suitable for most designs, reducing efficiency and limiting creativity. This paper implemented the Transformer model and conditional variational autoencoder (CVAE) to the graphic design layout generation task. It proposed an end-to-end graphic design layout generation model named LayoutT-CVAE. We also proposed element disentanglement and feature-based disentanglement strategies and introduce new graphic design principles and similarity metrics into the model, which significantly increased the controllability and interpretability of the deep model. Compared with the existing state-of-art models, the layout generated by ours performs better on many metrics.

(0 min) Binary Neural Network (BNN) method is an extreme application of convolutional neural network (CNN) parameter quantization. As opposed to the original CNN methods which employed floating-point computation with full-precision weights and activations, BBN uses 1-bit activations and weights. With BBNs, a significant amount of storage, network complexity and energy consumption can be reduced, and neural networks can be implemented more efficiently in embedded applications. Unfortunately, binarization causes severe information loss. A gap still exists between full-precision CNN models and their binarized counterparts. The recent developments in BNN have led to a lot of algorithms and solutions that have helped address this issue. This article provides a full overview of recent developments in BNN. The present paper focuses exclusively on 1-bit activations and weights networks, as opposed to previous surveys in which low-bit works are mixed in. In this paper, we conduct a complete investigation of BNN's development from their predecessors to the latest BNN algorithms and techniques, presenting a broad design pipeline, and discussing each module's variants. Along the way, this paper examines BNN (a) purpose: their early successes and challenges; (b) BNN optimization: selected representative works that contain key optimization techniques; (c) deployment: open-source frameworks for BNN modeling and development; (d) terminal: efficient computing architectures and devices for BNN and (e) applications: diverse applications with BNN. Moreover, this paper discusses potential directions and future research opportunities for the latest BNN algorithms and techniques, presents a broad design pipeline, and discusses each module's variants.

(0 min) UNet-based methods have shown outstanding performance in salient object detection (SOD), but are problematic in two aspects. 1) Indiscriminately integrating the encoder feature, which contains spatial information for multiple objects, and the decoder feature, which contains global information of the salient object, is likely to convey unnecessary details of non-salient objects to the decoder, hindering saliency detection. 2) To deal with ambiguous object boundaries and generate accurate saliency maps, the model needs additional branches, such as edge reconstructions, which leads to increasing computational cost. To address the problems, we propose a context fusion decoder network (CFDN) and near edge weighted loss (NEWLoss) function. The CFDN creates an accurate saliency map by integrating global context information and thus suppressing the influence of the unnecessary spatial information. NEWLoss accelerates learning of obscure boundaries without additional modules by generating weight maps on object boundaries. Our method is evaluated on four benchmarks and achieves state-of-the-art performance. We prove the effectiveness of the proposed method through comparative experiments.

(0 min) Quantization is a widely used technique to compress and accelerate deep neural networks. However, conventional quantization methods use the same bit-width for all (or most of) the layers, which often suffer significant accuracy degradation in the ultra-low precision regime and ignore the fact that emergent hardware accelerators begin to support mixed-precision computation. Consequently, we present a novel and principled framework to solve the mixed-precision quantization problem in this paper. Briefly speaking, we first formulate the mixed-precision quantization as a discrete constrained optimization problem. Then, to make the optimization tractable, we approximate the objective function with second-order Taylor expansion and propose an efficient approach to compute its Hessian matrix. Finally, based on the above simplification, we show that the original problem can be reformulated as a Multiple-Choice Knapsack Problem (MCKP) and propose a greedy search algorithm to solve it efficiently. Compared with existing mixed-precision quantization works, our method is derived in a principled way and much more computationally efficient. Moreover, extensive experiments conducted on the ImageNet dataset and various kinds of network architectures also demonstrate its superiority over existing uniform and mixed-precision quantization approaches.

(0 min) Oriented object detection in remote sensing images has made great progress in recent years. However, most of the current methods only focus on detecting targets, and cannot distinguish fine-grained objects well in complex scenes. In this technical report, we analyzed the key issues of fine-grained object recognition, and use an oriented feature alignment network (OFA-Net) to achieve high-performance fine-grained oriented object recognition in optical remote sensing images. OFA-Net achieves accurate object localization through a rotated bounding boxes refinement module. On this basis, the boundary-constrained rotation feature alignment module is applied to achieve local feature extraction, which is beneficial to fine-grained object classification. The single model of our method achieved mAP of 46.51\% in the GaoFen competition and won 3rd place in the ISPRS benchmark with the mAP of 43.73\%.

(0 min) Image fusion methods and metrics for their evaluation have conventionally used pixel-based or low-level features. However, for many applications, the aim of image fusion is to effectively combine the semantic content of the input images. This paper proposes a novel system for the semantic combination of visual content using pre-trained CNN network architectures. Our proposed semantic fusion is initiated through the fusion of the top layer feature map outputs (for each input image)through gradient updating of the fused image input (so-called image optimisation). Simple "choose maximum" and "local majority" filter based fusion rules are utilised for feature map fusion. This provides a simple method to combine layer outputs and thus a unique framework to fuse single-channel and colour images within a decomposition pre-trained for classification and therefore aligned with semantic fusion. Furthermore, class activation mappings of each input image are used to combine semantic information at a higher level. The developed methods are able to give equivalent low-level fusion performance to state of the art methods while providing a unique architecture to combine semantic information from multiple images.

(0 min) Color modelling and extraction is an important topic in fashion, art, and design. Recommender systems, color-based retrieval, decorating, and fashion design can benefit from color extraction tools. Research has shown that modeling color so that it can be automatically analyzed and / or extracted is a difficult task. Unlike machines, color perception, although very subjective, is much simpler for humans. That being said, the first step in color modeling is to estimate the number of colors in the item / object. This is because color models can take advantage of the number of colors as the seed for better modelling, e.g., to make color extraction further deterministic. We aim in this work to develop and test models that can count the number of colors of clothing and other items. We propose a novel color counting method based on cumulative color histogram, which stands out among other methods. We compare the method we propose with other methods that utilize exhaustive color search that uses Gaussian Mixture Models (GMMs) and K-Means as bases for scoring the optimal number of colors, in addition to another method that relies on deep learning models. Unfortunately, the GMM, K-Means, and Deep Learning models all fail to accurately capture the number of colors. Our proposed method can provide the color baseline that can be used in AI-based fashion applications, and can also find applications in other areas, for example, interior design. To the best of our knowledge, this work is the first of its kind that addresses the problem of color-counting machine.

(0 min) The required amount of labeled data is one of the biggest issues in deep learning. Semi-Supervised Learning can potentially solve this issue by using additional unlabeled data. However, many datasets suffer from variability in the annotations. The aggregated labels from these annotation are not consistent between different annotators and thus are considered fuzzy. These fuzzy labels are often not considered by Semi-Supervised Learning. This leads either to an inferior performance or to higher initial annotation costs in the complete machine learning development cycle. We envision the incorporation of fuzzy labels into Semi-Supervised Learning and give a proof-of-concept of the potential lower costs and higher consistency in the complete development cycle. As part of our concept, we discuss current limitations, futures research opportunities and potential broad impacts.

(0 min) Video captioning is a challenging task since it requires generating sentences describing various diverse and complex videos. Existing video captioning models lack adequate visual representation due to the neglect of the existence of gaps between videos and texts. To bridge this gap, in this paper, we propose a CLIP4Caption framework that improves video captioning based on a CLIP-enhanced video-text matching network (VTM). This framework is taking full advantage of the information from both vision and language and enforcing the model to learn strongly text-correlated video features for text generation. Besides, unlike most existing models using LSTM or GRU as the sentence decoder, we adopt a Transformer structured decoder network to effectively learn the long-range visual and language dependency. Additionally, we introduce a novel ensemble strategy for captioning tasks. Experimental results demonstrate the effectiveness of our method on two datasets: 1) on MSR-VTT dataset, our method achieved a new state-of-the-art result with a significant gain of up to 10% in CIDEr; 2) on the private test data, our method ranking 2nd place in the ACM MM multimedia grand challenge 2021: Pre-training for Video Understanding Challenge. It is noted that our model is only trained on the MSR-VTT dataset.

(0 min) The VALUE (Video-And-Language Understanding Evaluation) benchmark is newly introduced to evaluate and analyze multi-modal representation learning algorithms on three video-and-language tasks: Retrieval, QA, and Captioning. The main objective of the VALUE challenge is to train a task-agnostic model that is simultaneously applicable for various tasks with different characteristics. This technical report describes our winning strategies for the VALUE challenge: 1) single model optimization, 2) transfer learning with visual concepts, and 3) task-aware ensemble. The first and third strategies are designed to address heterogeneous characteristics of each task, and the second one is to leverage rich and fine-grained visual information. We provide a detailed and comprehensive analysis with extensive experimental results. Based on our approach, we ranked first place on the VALUE and QA phases for the competition.

(0 min) Domain adaptive semantic segmentation is recognized as a promising technique to alleviate the domain shift between the labeled source domain and the unlabeled target domain in many real-world applications, such as automatic pilot. However, large amounts of source domain data often introduce significant costs in storage and training, and sometimes the source data is inaccessible due to privacy policies. To address these problems, we investigate domain adaptive semantic segmentation without source data, which assumes that the model is pre-trained on the source domain, and then adapting to the target domain without accessing source data anymore. Since there is no supervision from the source domain data, many self-training methods tend to fall into the ``winner-takes-all'' dilemma, where the {\it majority} classes totally dominate the segmentation networks and the networks fail to classify the {\it minority} classes. Consequently, we propose an effective framework for this challenging problem with two components: positive learning and negative learning. In positive learning, we select the class-balanced pseudo-labeled pixels with intra-class threshold, while in negative learning, for each pixel, we investigate which category the pixel does not belong to with the proposed heuristic complementary label selection. Notably, our framework can be easily implemented and incorporated with other methods to further enhance the performance. Extensive experiments on two widely-used synthetic-to-real benchmarks demonstrate our claims and the effectiveness of our framework, which outperforms the baseline with a large margin. Code is available at \url{https://github.com/fumyou13/LDBE}.

(0 min) Computational modeling of the emotions evoked by art in humans is a challenging problem because of the subjective and nuanced nature of art and affective signals. In this paper, we consider the above-mentioned problem of understanding emotions evoked in viewers by artwork using both text and visual modalities. Specifically, we analyze images and the accompanying text captions from the viewers expressing emotions as a multimodal classification task. Our results show that single-stream multimodal transformer-based models like MMBT and VisualBERT perform better compared to both image-only models and dual-stream multimodal models having separate pathways for text and image modalities. We also observe improvements in performance for extreme positive and negative emotion classes, when a single-stream model like MMBT is compared with a text-only transformer model like BERT.

(0 min) Deep Learning (DL) requires a large amount of training data to provide quality outcomes. However, the field of medical imaging suffers from the lack of sufficient data for properly training DL models because medical images require manual labelling carried out by clinical experts thus the process is time-consuming, expensive, and error-prone. Recently, transfer learning (TL) was introduced to reduce the need for the annotation procedure by means of transferring the knowledge performed by a previous task and then fine-tuning the result using a relatively small dataset. Nowadays, multiple classification methods from medical imaging make use of TL from general-purpose pre-trained models, e.g., ImageNet, which has been proven to be ineffective due to the mismatch between the features learned from natural images (ImageNet) and those more specific from medical images especially medical gray images such as X-rays. ImageNet does not have grayscale images such as MRI, CT, and X-ray. In this paper, we propose a novel DL model to be used for addressing classification tasks of medical imaging, called MedNet. To do so, we aim to issue two versions of MedNet. The first one is Gray-MedNet which will be trained on 3M publicly available gray-scale medical images including MRI, CT, X-ray, ultrasound, and PET. The second version is Color-MedNet which will be trained on 3M publicly available color medical images including histopathology, taken images, and many others. To validate the effectiveness MedNet, both versions will be fine-tuned to train on the target tasks of a more reduced set of medical images. MedNet performs as the pre-trained model to tackle any real-world application from medical imaging and achieve the level of generalization needed for dealing with medical imaging tasks, e.g. classification. MedNet would serve the research community as a baseline for future research.

(0 min) Eliminating the covariate shift cross domains is one of the common methods to deal with the issue of domain shift in visual unsupervised domain adaptation. However, current alignment methods, especially the prototype based or sample-level based methods neglect the structural properties of the underlying distribution and even break the condition of covariate shift. To relieve the limitations and conflicts, we introduce a novel concept named (virtual) mirror, which represents the equivalent sample in another domain. The equivalent sample pairs, named mirror pairs reflect the natural correspondence of the empirical distributions. Then a mirror loss, which aligns the mirror pairs cross domains, is constructed to enhance the alignment of the domains. The proposed method does not distort the internal structure of the underlying distribution. We also provide theoretical proof that the mirror samples and mirror loss have better asymptotic properties in reducing the domain shift. By applying the virtual mirror and mirror loss to the generic unsupervised domain adaptation model, we achieved consistent superior performance on several mainstream benchmarks.

(0 min) Weakly supervised semantic segmentation produces pixel-level localization from class labels; however, a classifier trained on such labels is likely to focus on a small discriminative region of the target object. We interpret this phenomenon using the information bottleneck principle: the final layer of a deep neural network, activated by the sigmoid or softmax activation functions, causes an information bottleneck, and as a result, only a subset of the task-relevant information is passed on to the output. We first support this argument through a simulated toy experiment and then propose a method to reduce the information bottleneck by removing the last activation function. In addition, we introduce a new pooling method that further encourages the transmission of information from non-discriminative regions to the classification. Our experimental evaluations demonstrate that this simple modification significantly improves the quality of localization maps on both the PASCAL VOC 2012 and MS COCO 2014 datasets, exhibiting a new state-of-the-art performance for weakly supervised semantic segmentation. The code is available at: https://github.com/jbeomlee93/RIB.

(0 min) Learning generative object models from unlabelled videos is a long standing problem and required for causal scene modeling. We decompose this problem into three easier subtasks, and provide candidate solutions for each of them. Inspired by the Common Fate Principle of Gestalt Psychology, we first extract (noisy) masks of moving objects via unsupervised motion segmentation. Second, generative models are trained on the masks of the background and the moving objects, respectively. Third, background and foreground models are combined in a conditional "dead leaves" scene model to sample novel scene configurations where occlusions and depth layering arise naturally. To evaluate the individual stages, we introduce the Fishbowl dataset positioned between complex real-world scenes and common object-centric benchmarks of simplistic objects. We show that our approach allows learning generative models that generalize beyond the occlusions present in the input videos, and represent scenes in a modular fashion that allows sampling plausible scenes outside the training distribution by permitting, for instance, object numbers or densities not observed in the training set.

(0 min) In deep multi-view stereo networks, cost regularization is crucial to achieve accurate depth estimation. Since 3D cost volume filtering is usually memory-consuming, recurrent 2D cost map regularization has recently become popular and has shown great potential in reconstructing 3D models of different scales. However, existing recurrent methods only model the local dependencies in the depth domain, which greatly limits the capability of capturing the global scene context along the depth dimension. To tackle this limitation, we propose a novel non-local recurrent regularization network for multi-view stereo, named NR2-Net. Specifically, we design a depth attention module to capture non-local depth interactions within a sliding depth block. Then, the global scene context between different blocks is modeled in a gated recurrent manner. This way, the long-range dependencies along the depth dimension are captured to facilitate the cost regularization. Moreover, we design a dynamic depth map fusion strategy to improve the algorithm robustness. Our method achieves state-of-the-art reconstruction results on both DTU and Tanks and Temples datasets.

(0 min) Modern neural network architectures can leverage large amounts of data to generalize well within the training distribution. However, they are less capable of systematic generalization to data drawn from unseen but related distributions, a feat that is hypothesized to require compositional reasoning and reuse of knowledge. In this work, we present Neural Interpreters, an architecture that factorizes inference in a self-attention network as a system of modules, which we call \emph{functions}. Inputs to the model are routed through a sequence of functions in a way that is end-to-end learned. The proposed architecture can flexibly compose computation along width and depth, and lends itself well to capacity extension after training. To demonstrate the versatility of Neural Interpreters, we evaluate it in two distinct settings: image classification and visual abstract reasoning on Raven Progressive Matrices. In the former, we show that Neural Interpreters perform on par with the vision transformer using fewer parameters, while being transferrable to a new task in a sample efficient manner. In the latter, we find that Neural Interpreters are competitive with respect to the state-of-the-art in terms of systematic generalization

(0 min) In this paper, we discuss the initial attempts at boosting understanding human language based on deep-learning models with quantum computing. We successfully train a quantum-enhanced Long Short-Term Memory network to perform the parts-of-speech tagging task via numerical simulations. Moreover, a quantum-enhanced Transformer is proposed to perform the sentiment analysis based on the existing dataset.

(0 min) In multimodal assistant, where vision is also one of the input modalities, the identification of user intent becomes a challenging task as visual input can influence the outcome. Current digital assistants take spoken input and try to determine the user intent from conversational or device context. So, a dataset, which includes visual input (i.e. images or videos for the corresponding questions targeted for multimodal assistant use cases, is not readily available. The research in visual question answering (VQA) and visual question generation (VQG) is a great step forward. However, they do not capture questions that a visually-abled person would ask multimodal assistants. Moreover, many times questions do not seek information from external knowledge. In this paper, we provide a new dataset, MMIU (MultiModal Intent Understanding), that contains questions and corresponding intents provided by human annotators while looking at images. We, then, use this dataset for intent classification task in multimodal digital assistant. We also experiment with various approaches for combining vision and language features including the use of multimodal transformer for classification of image-question pairs into 14 intents. We provide the benchmark results and discuss the role of visual and text features for the intent classification task on our dataset.

(0 min) Ultrasound imaging is a commonly used technology for visualising patient anatomy in real-time during diagnostic and therapeutic procedures. High operator dependency and low reproducibility make ultrasound imaging and interpretation challenging with a steep learning curve. Automatic image classification using deep learning has the potential to overcome some of these challenges by supporting ultrasound training in novices, as well as aiding ultrasound image interpretation in patient with complex pathology for more experienced practitioners. However, the use of deep learning methods requires a large amount of data in order to provide accurate results. Labelling large ultrasound datasets is a challenging task because labels are retrospectively assigned to 2D images without the 3D spatial context available in vivo or that would be inferred while visually tracking structures between frames during the procedure. In this work, we propose a multi-modal convolutional neural network (CNN) architecture that labels endoscopic ultrasound (EUS) images from raw verbal comments provided by a clinician during the procedure. We use a CNN composed of two branches, one for voice data and another for image data, which are joined to predict image labels from the spoken names of anatomical landmarks. The network was trained using recorded verbal comments from expert operators. Our results show a prediction accuracy of 76% at image level on a dataset with 5 different labels. We conclude that the addition of spoken commentaries can increase the performance of ultrasound image classification, and eliminate the burden of manually labelling large EUS datasets necessary for deep learning applications.

(0 min) We propose a novel approach to translate unpaired contrast computed tomography (CT) scans to non-contrast CT scans and the other way around. Solving this task has two important applications: (i) to automatically generate contrast CT scans for patients for whom injecting contrast substance is not an option, and (ii) to enhance alignment between contrast and non-contrast CT by reducing the differences induced by the contrast substance before registration. Our approach is based on cycle-consistent generative adversarial convolutional transformers, for short, CyTran. Our neural model can be trained on unpaired images, due to the integration of a cycle-consistency loss. To deal with high-resolution images, we design a hybrid architecture based on convolutional and multi-head attention layers. In addition, we introduce a novel data set, Coltea-Lung-CT-100W, containing 3D triphasic lung CT scans (with a total of 37,290 images) collected from 100 female patients. Each scan contains three phases (non-contrast, early portal venous, and late arterial), allowing us to perform experiments to compare our novel approach with state-of-the-art methods for image style transfer. Our empirical results show that CyTran outperforms all competing methods. Moreover, we show that CyTran can be employed as a preliminary step to improve a state-of-the-art medical image alignment method. We release our novel model and data set as open source at: https://github.com/ristea/cycle-transformer.

(0 min) Recommendation systems play an important role in today's digital world. They have found applications in various applications such as music platforms, e.g., Spotify, and movie streaming services, e.g., Netflix. Less research effort has been devoted to physical exercise recommendation systems. Sedentary lifestyles have become the major driver of several diseases as well as healthcare costs. In this paper, we develop a recommendation system for daily exercise activities to users based on their history, profile and similar users. The developed recommendation system uses a deep recurrent neural network with user-profile attention and temporal attention mechanisms. Moreover, exercise recommendation systems are significantly different from streaming recommendation systems in that we are not able to collect click feedback from the participants in exercise recommendation systems. Thus, we propose a real-time, expert-in-the-loop active learning procedure. The active learners calculate the uncertainty of the recommender at each time step for each user and ask an expert for a recommendation when the certainty is low. In this paper, we derive the probability distribution function of marginal distance, and use it to determine when to ask experts for feedback. Our experimental results on a mHealth dataset show improved accuracy after incorporating the real-time active learner with the recommendation system.

(0 min) Lesion images are frequently taken in open-set settings. Because of this, the image data generated is extremely varied in nature.It is difficult for a convolutional neural network to find proper features and generalise well, as a result content based image retrieval (CBIR) system for lesion images are difficult to build. This paper explores this domain and proposes multiple similarity measures which uses Style Loss and Dice Coefficient via a novel similarity measure called I1-Score. Out of the CBIR similarity measures proposed, pure style loss approach achieves a remarkable accuracy increase over traditional approaches like Euclidean Distance and Cosine Similarity. The I1-Scores using style loss performed better than traditional approaches by a small margin, whereas, I1-Scores with dice-coefficient faired very poorly. The model used is trained using ensemble learning for better generalization.

(0 min) Deep neural networks can be roughly divided into deterministic neural networks and stochastic neural networks.The former is usually trained to achieve a mapping from input space to output space via maximum likelihood estimation for the weights, which leads to deterministic predictions during testing. In this way, a specific weights set is estimated while ignoring any uncertainty that may occur in the proper weight space. The latter introduces randomness into the framework, either by assuming a prior distribution over model parameters (i.e. Bayesian Neural Networks) or including latent variables (i.e. generative models) to explore the contribution of latent variables for model predictions, leading to stochastic predictions during testing. Different from the former that achieves point estimation, the latter aims to estimate the prediction distribution, making it possible to estimate uncertainty, representing model ignorance about its predictions. We claim that conventional deterministic neural network based dense prediction tasks are prone to overfitting, leading to over-confident predictions, which is undesirable for decision making. In this paper, we investigate stochastic neural networks and uncertainty estimation techniques to achieve both accurate deterministic prediction and reliable uncertainty estimation. Specifically, we work on two types of uncertainty estimations solutions, namely ensemble based methods and generative model based methods, and explain their pros and cons while using them in fully/semi/weakly-supervised framework. Due to the close connection between uncertainty estimation and model calibration, we also introduce how uncertainty estimation can be used for deep model calibration to achieve well-calibrated models, namely dense model calibration. Code and data are available at https://github.com/JingZhang617/UncertaintyEstimation.

(0 min) Graph Convolutional Representation (GCR) has achieved impressive performance for graph data representation. However, existing GCR is generally defined on the input fixed graph which may restrict the representation capacity and also be vulnerable to the structural attacks and noises. To address this issue, we propose a novel Latent Graph Convolutional Representation (LatGCR) for robust graph data representation and learning. Our LatGCR is derived based on reformulating graph convolutional representation from the aspect of graph neighborhood reconstruction. Given an input graph $\textbf{A}$, LatGCR aims to generate a flexible latent graph $\widetilde{\textbf{A}}$ for graph convolutional representation which obviously enhances the representation capacity and also performs robustly w.r.t graph structural attacks and noises. Moreover, LatGCR is implemented in a self-supervised manner and thus provides a basic block for both supervised and unsupervised graph learning tasks. Experiments on several datasets demonstrate the effectiveness and robustness of LatGCR.

(0 min) We introduce a framework for multi-camera 3D object detection. In contrast to existing works, which estimate 3D bounding boxes directly from monocular images or use depth prediction networks to generate input for 3D object detection from 2D information, our method manipulates predictions directly in 3D space. Our architecture extracts 2D features from multiple camera images and then uses a sparse set of 3D object queries to index into these 2D features, linking 3D positions to multi-view images using camera transformation matrices. Finally, our model makes a bounding box prediction per object query, using a set-to-set loss to measure the discrepancy between the ground-truth and the prediction. This top-down approach outperforms its bottom-up counterpart in which object bounding box prediction follows per-pixel depth estimation, since it does not suffer from the compounding error introduced by a depth prediction model. Moreover, our method does not require post-processing such as non-maximum suppression, dramatically improving inference speed. We achieve state-of-the-art performance on the nuScenes autonomous driving benchmark.

(0 min) While deep reinforcement learning (RL) promises freedom from hand-labeled data, great successes, especially for Embodied AI, require significant work to create supervision via carefully shaped rewards. Indeed, without shaped rewards, i.e., with only terminal rewards, present-day Embodied AI results degrade significantly across Embodied AI problems from single-agent Habitat-based PointGoal Navigation (SPL drops from 55 to 0) and two-agent AI2-THOR-based Furniture Moving (success drops from 58% to 1%) to three-agent Google Football-based 3 vs. 1 with Keeper (game score drops from 0.6 to 0.1). As training from shaped rewards doesn't scale to more realistic tasks, the community needs to improve the success of training with terminal rewards. For this we propose GridToPix: 1) train agents with terminal rewards in gridworlds that generically mirror Embodied AI environments, i.e., they are independent of the task; 2) distill the learned policy into agents that reside in complex visual worlds. Despite learning from only terminal rewards with identical models and RL algorithms, GridToPix significantly improves results across tasks: from PointGoal Navigation (SPL improves from 0 to 64) and Furniture Moving (success improves from 1% to 25%) to football gameplay (game score improves from 0.1 to 0.6). GridToPix even helps to improve the results of shaped reward training.

(0 min) Persistent Homologies have been successfully used to increase the performance of deep networks trained to detect curvilinear structures and to improve the topological quality of the results. However, existing methods are very global and ignore the location of topological features. In this paper, we introduce an approach that relies on a new filtration function to account for location during network training. We demonstrate experimentally on 2D images of roads and 3D image stacks of neuronal processes that networks trained in this manner are better at recovering the topology of the curvilinear structures they extract.

(0 min) Pairwise dot product-based attention allows Transformers to exchange information between tokens in an input-dependent way, and is key to their success across diverse applications in language and vision. However, a typical Transformer model computes such pairwise attention scores repeatedly for the same sequence, in multiple heads in multiple layers. We systematically analyze the empirical similarity of these scores across heads and layers and find them to be considerably redundant, especially adjacent layers showing high similarity. Motivated by these findings, we propose a novel architecture that reuses attention scores computed in one layer in multiple subsequent layers. Experiments on a number of standard benchmarks show that reusing attention delivers performance equivalent to or better than standard transformers, while reducing both compute and memory usage.

(0 min) The COVID-19 pandemic raises the problem of adapting face recognition systems to the new reality, where people may wear surgical masks to cover their noses and mouths. Traditional data sets (e.g., CelebA, CASIA-WebFace) used for training these systems were released before the pandemic, so they now seem unsuited due to the lack of examples of people wearing masks. We propose a method for enhancing data sets containing faces without masks by creating synthetic masks and overlaying them on faces in the original images. Our method relies on SparkAR Studio, a developer program made by Facebook that is used to create Instagram face filters. In our approach, we use 9 masks of different colors, shapes and fabrics. We employ our method to generate a number of 445,446 (90%) samples of masks for the CASIA-WebFace data set and 196,254 (96.8%) masks for the CelebA data set, releasing the mask images at https://github.com/securifai/masked_faces. We show that our method produces significantly more realistic training examples of masks overlaid on faces by asking volunteers to qualitatively compare it to other methods or data sets designed for the same task. We also demonstrate the usefulness of our method by evaluating state-of-the-art face recognition systems (FaceNet, VGG-face, ArcFace) trained on our enhanced data sets and showing that they outperform equivalent systems trained on original data sets (containing faces without masks) or competing data sets (containing masks generated by related methods), when the test benchmarks contain masked faces.

(2 min) Recent progress in fine-grained gesture and action classification, and machine translation, point to the possibility of automated sign language recognition becoming a reality. A key stumbling block in making progress towards this goal is a lack of appropriate training data, stemming from the high complexity of sign annotation and a limited supply of qualified annotators. In this work, we introduce a new scalable approach to data collection for sign recognition in continuous videos. We make use of weakly-aligned subtitles for broadcast footage together with a keyword spotting method to automatically localise sign-instances for a vocabulary of 1,000 signs in 1,000 hours of video. We make the following contributions: (1) We show how to use mouthing cues from signers to obtain high-quality annotations from video data - the result is the BSL-1K dataset, a collection of British Sign Language (BSL) signs of unprecedented scale; (2) We show that we can use BSL-1K to train strong sign recognition models for co-articulated signs in BSL and that these models additionally form excellent pretraining for other sign languages and benchmarks - we exceed the state of the art on both the MSASL and WLASL benchmarks. Finally, (3) we propose new large-scale evaluation sets for the tasks of sign recognition and sign spotting and provide baselines which we hope will serve to stimulate research in this area.

(2 min) Current state-of-the-art approaches for image captioning typically adopt an autoregressive manner, i.e., generating descriptions word by word, which suffers from slow decoding issue and becomes a bottleneck in real-time applications. Non-autoregressive image captioning with continuous iterative refinement, which eliminates the sequential dependence in a sentence generation, can achieve comparable performance to the autoregressive counterparts with a considerable acceleration. Nevertheless, based on a well-designed experiment, we empirically proved that iteration times can be effectively reduced when providing sufficient prior knowledge for the language decoder. Towards that end, we propose a novel two-stage framework, referred to as Semi-Autoregressive Image Captioning (SAIC), to make a better trade-off between performance and speed. The proposed SAIC model maintains autoregressive property in global but relieves it in local. Specifically, SAIC model first jumpily generates an intermittent sequence in an autoregressive manner, that is, it predicts the first word in every word group in order. Then, with the help of the partially deterministic prior information and image features, SAIC model non-autoregressively fills all the skipped words with one iteration. Experimental results on the MS COCO benchmark demonstrate that our SAIC model outperforms the preceding non-autoregressive image captioning models while obtaining a competitive inference speedup. Code is available at https://github.com/feizc/SAIC.

(2 min) Mainstream approaches for unsupervised domain adaptation (UDA) learn domain-invariant representations to narrow the domain shift. Recently, self-training has been gaining momentum in UDA, which exploits unlabeled target data by training with target pseudo-labels. However, as corroborated in this work, under distributional shift in UDA, the pseudo-labels can be unreliable in terms of their large discrepancy from target ground truth. Thereby, we propose Cycle Self-Training (CST), a principled self-training algorithm that explicitly enforces pseudo-labels to generalize across domains. CST cycles between a forward step and a reverse step until convergence. In the forward step, CST generates target pseudo-labels with a source-trained classifier. In the reverse step, CST trains a target classifier using target pseudo-labels, and then updates the shared representations to make the target classifier perform well on the source data. We introduce the Tsallis entropy as a confidence-friendly regularization to improve the quality of target pseudo-labels. We analyze CST theoretically under realistic assumptions, and provide hard cases where CST recovers target ground truth, while both invariant feature learning and vanilla self-training fail. Empirical results indicate that CST significantly improves over the state-of-the-arts on visual recognition and sentiment analysis benchmarks.

(2 min) Predicting vulnerable road user behavior is an essential prerequisite for deploying Automated Driving Systems (ADS) in the real-world. Pedestrian crossing intention should be recognized in real-time, especially for urban driving. Recent works have shown the potential of using vision-based deep neural network models for this task. However, these models are not robust and certain issues still need to be resolved. First, the global spatio-temproal context that accounts for the interaction between the target pedestrian and the scene has not been properly utilized. Second, the optimum strategy for fusing different sensor data has not been thoroughly investigated. This work addresses the above limitations by introducing a novel neural network architecture to fuse inherently different spatio-temporal features for pedestrian crossing intention prediction. We fuse different phenomena such as sequences of RGB imagery, semantic segmentation masks, and ego-vehicle speed in an optimum way using attention mechanisms and a stack of recurrent neural networks. The optimum architecture was obtained through exhaustive ablation and comparison studies. Extensive comparative experiments on the JAAD pedestrian action prediction benchmark demonstrate the effectiveness of the proposed method, where state-of-the-art performance was achieved. Our code is open-source and publicly available.

(2 min) Ship detection in remote sensing images plays a crucial role in various applications and has drawn increasing attention in recent years. However, existing arbitrary-oriented ship detection methods are generally developed on a set of predefined rotated anchor boxes. These predefined boxes not only lead to inaccurate angle predictions but also introduce extra hyper-parameters and high computational cost. Moreover, the prior knowledge of ship size has not been fully exploited by existing methods, which hinders the improvement of their detection accuracy. Aiming at solving the above issues, in this paper, we propose a center-head point extraction based detector (named CHPDet) to achieve arbitrary-oriented ship detection in remote sensing images. Our CHPDet formulates arbitrary-oriented ships as rotated boxes with head points which are used to determine the direction. And rotated Gaussian kernel is used to map the annotations into target heatmaps. Keypoint estimation is performed to find the center of ships. Then, the size and head point of the ships are regressed. The orientation-invariant model (OIM) is also used to produce orientation-invariant feature maps. Finally, we use the target size as prior to finetune the results. Moreover, we introduce a new dataset for multi-class arbitrary-oriented ship detection in remote sensing images at a fixed ground sample distance (GSD) which is named FGSD2021. Experimental results on FGSD2021 and two other widely used data sets, i.e., HRSC2016, and UCAS-AOD demonstrate that our CHPDet achieves state-of-the-art performance and can well distinguish between bow and stern. Code and FGSD2021 dataset are available at https://github.com/zf020114/CHPDet.

(2 min) Deep learning has shown remarkable progress in a wide range of problems. However, efficient training of such models requires large-scale datasets, and getting annotations for such datasets can be challenging and costly. In this work, we explore the use of user-generated freely available labels from web videos for video understanding. We create a benchmark dataset consisting of around 2 million videos with associated user-generated annotations and other meta information. We utilize the collected dataset for action classification and demonstrate its usefulness with existing small-scale annotated datasets, UCF101 and HMDB51. We study different loss functions and two pretraining strategies, simple and self-supervised learning. We also show how a network pretrained on the proposed dataset can help against video corruption and label noise in downstream datasets. We present this as a benchmark dataset in noisy learning for video understanding. The dataset, code, and trained models will be publicly available for future research.

(3 min) Joint registration of a stack of 2D histological sections to recover 3D structure (``3D histology reconstruction'') finds application in areas such as atlas building and validation of \emph{in vivo} imaging. Straightforward pairwise registration of neighbouring sections yields smooth reconstructions but has well-known problems such as ``banana effect'' (straightening of curved structures) and ``z-shift'' (drift). While these problems can be alleviated with an external, linearly aligned reference (e.g., Magnetic Resonance (MR) images), registration is often inaccurate due to contrast differences and the strong nonlinear distortion of the tissue, including artefacts such as folds and tears. In this paper, we present a probabilistic model of spatial deformation that yields reconstructions for multiple histological stains that that are jointly smooth, robust to outliers, and follow the reference shape. The model relies on a spanning tree of latent transforms connecting all the sections and slices of the reference volume, and assumes that the registration between any pair of images can be see as a noisy version of the composition of (possibly inverted) latent transforms connecting the two images. Bayesian inference is used to compute the most likely latent transforms given a set of pairwise registrations between image pairs within and across modalities. The framework is used for accurate 3D reconstruction of two stains (Nissl and parvalbumin) from the Allen human brain atlas, showing its benefits on real data with severe distortions. Moreover, we also provide the registration of the reconstructed volume to MNI space, bridging the gaps between two of the most widely used atlases in histology and MRI. The 3D reconstructed volumes and atlas registration can be downloaded from https://openneuro.org/datasets/ds003590. The code is freely available at https://github.com/acasamitjana/3dhirest.

(3 min) The latest Deep Learning (DL) models for detection and classification have achieved an unprecedented performance over classical machine learning algorithms. However, DL models are black-box methods hard to debug, interpret, and certify. DL alone cannot provide explanations that can be validated by a non technical audience. In contrast, symbolic AI systems that convert concepts into rules or symbols -- such as knowledge graphs -- are easier to explain. However, they present lower generalisation and scaling capabilities. A very important challenge is to fuse DL representations with expert knowledge. One way to address this challenge, as well as the performance-explainability trade-off is by leveraging the best of both streams without obviating domain expert knowledge. We tackle such problem by considering the symbolic knowledge is expressed in form of a domain expert knowledge graph. We present the eXplainable Neural-symbolic learning (X-NeSyL) methodology, designed to learn both symbolic and deep representations, together with an explainability metric to assess the level of alignment of machine and human expert explanations. The ultimate objective is to fuse DL representations with expert domain knowledge during the learning process to serve as a sound basis for explainability. X-NeSyL methodology involves the concrete use of two notions of explanation at inference and training time respectively: 1) EXPLANet: Expert-aligned eXplainable Part-based cLAssifier NETwork Architecture, a compositional CNN that makes use of symbolic representations, and 2) SHAP-Backprop, an explainable AI-informed training procedure that guides the DL process to align with such symbolic representations in form of knowledge graphs. We showcase X-NeSyL methodology using MonuMAI dataset for monument facade image classification, and demonstrate that our approach improves explainability and performance.

(2 min) During the COVID-19 pandemic, most of the human-to-human interactions have been stopped. To mitigate the spread of deadly coronavirus, many offices took the initiative so that the employees can work from home. But, tracking the employees and finding out if they are really performing what they were supposed to turn out to be a serious challenge for all the companies and organizations who are facilitating "Work From Home". To deal with the challenge effectively, we came up with a solution to track the employees with face recognition. We have been testing this system experimentally for our office. To train the face recognition module, we used FaceNet with KNN using the Labeled Faces in the Wild (LFW) dataset and achieved 97.8\% accuracy. We integrated the trained model into our central system, where the employees log their time. In this paper, we discuss in brief the system we have been experimenting with and the pros and cons of the system.

(2 min) In video data, busy motion details from moving regions are conveyed within a specific frequency bandwidth in the frequency domain. Meanwhile, the rest of the frequencies of video data are encoded with quiet information with substantial redundancy, which causes low processing efficiency in existing video models that take as input raw RGB frames. In this paper, we consider allocating intenser computation for the processing of the important busy information and less computation for that of the quiet information. We design a trainable Motion Band-Pass Module (MBPM) for separating busy information from quiet information in raw video data. By embedding the MBPM into a two-pathway CNN architecture, we define a Busy-Quiet Net (BQN). The efficiency of BQN is determined by avoiding redundancy in the feature space processed by the two pathways: one operating on Quiet features of low-resolution, while the other processes Busy features. The proposed BQN outperforms many recent video processing models on Something-Something V1, Kinetics400, UCF101 and HMDB51 datasets.

(2 min) Training deep neural networks may be challenging in real world data. Using models as black-boxes, even with transfer learning, can result in poor generalization or inconclusive results when it comes to small datasets or specific applications. This tutorial covers the basic steps as well as more recent options to improve models, in particular, but not restricted to, supervised learning. It can be particularly useful in datasets that are not as well-prepared as those in challenges, and also under scarce annotation and/or small data. We describe basic procedures: as data preparation, optimization and transfer learning, but also recent architectural choices such as use of transformer modules, alternative convolutional layers, activation functions, wide and deep networks, as well as training procedures including as curriculum, contrastive and self-supervised learning.

(2 min) When learning tasks over time, artificial neural networks suffer from a problem known as Catastrophic Forgetting (CF). This happens when the weights of a network are overwritten during the training of a new task causing forgetting of old information. To address this issue, we propose MetA Reusable Knowledge or MARK, a new method that fosters weight reusability instead of overwriting when learning a new task. Specifically, MARK keeps a set of shared weights among tasks. We envision these shared weights as a common Knowledge Base (KB) that is not only used to learn new tasks, but also enriched with new knowledge as the model learns new tasks. Key components behind MARK are two-fold. On the one hand, a metalearning approach provides the key mechanism to incrementally enrich the KB with new knowledge and to foster weight reusability among tasks. On the other hand, a set of trainable masks provides the key mechanism to selectively choose from the KB relevant weights to solve each task. By using MARK, we achieve state of the art results in several popular benchmarks, surpassing the best performing methods in terms of average accuracy by over 10% on the 20-Split-MiniImageNet dataset, while achieving almost zero forgetfulness using 55% of the number of parameters. Furthermore, an ablation study provides evidence that, indeed, MARK is learning reusable knowledge that is selectively used by each task.

(2 min) The phenomenon of Human Pose Estimation (HPE) is a problem that has been explored over the years, particularly in computer vision. But what exactly is it? To answer this, the concept of a pose must first be understood. Pose can be defined as the arrangement of human joints in a specific manner. Therefore, we can define the problem of Human Pose Estimation as the localization of human joints or predefined landmarks in images and videos. There are several types of pose estimation, including body, face, and hand, as well as many aspects to it. This paper will cover them, starting with the classical approaches to HPE to the Deep Learning based models.

(2 min) Face Anti-Spoofing (FAS) is essential to secure face recognition systems and has been extensively studied in recent years. Although deep neural networks (DNNs) for the FAS task have achieved promising results in intra-dataset experiments with similar distributions of training and testing data, the DNNs' generalization ability is limited under the cross-domain scenarios with different distributions of training and testing data. To improve the generalization ability, recent hybrid methods have been explored to extract task-aware handcrafted features (e.g., Local Binary Pattern) as discriminative information for the input of DNNs. However, the handcrafted feature extraction relies on experts' domain knowledge, and how to choose appropriate handcrafted features is underexplored. To this end, we propose a learnable network to extract Meta Pattern (MP) in our learning-to-learn framework. By replacing handcrafted features with the MP, the discriminative information from MP is capable of learning a more generalized model. Moreover, we devise a two-stream network to hierarchically fuse the input RGB image and the extracted MP by using our proposed Hierarchical Fusion Module (HFM). We conduct comprehensive experiments and show that our MP outperforms the compared handcrafted features. Also, our proposed method with HFM and the MP can achieve state-of-the-art performance on two different domain generalization evaluation benchmarks.

(2 min) We aim at advancing open-vocabulary object detection, which detects objects described by arbitrary text inputs. The fundamental challenge is the availability of training data. Existing object detection datasets only contain hundreds of categories, and it is costly to scale further. To overcome this challenge, we propose ViLD, a training method via Vision and Language knowledge Distillation. Our method distills the knowledge from a pretrained open-vocabulary image classification model (teacher) into a two-stage detector (student). Specifically, we use the teacher model to encode category texts and image regions of object proposals. Then we train a student detector, whose region embeddings of detected boxes are aligned with the text and image embeddings inferred by the teacher. We benchmark on LVIS by holding out all rare categories as novel categories not seen during training. ViLD obtains 16.1 mask AP$_r$, even outperforming the supervised counterpart by 3.8 with a ResNet-50 backbone. The model can directly transfer to other datasets without finetuning, achieving 72.2 AP$_{50}$, 36.6 AP and 11.8 AP on PASCAL VOC, COCO and Objects365, respectively. On COCO, ViLD outperforms previous SOTA by 4.8 on novel AP and 11.4 on overall AP.

(2 min) Recent years have witnessed an explosion of science conspiracy videos on the Internet, challenging science epistemology and public understanding of science. Scholars have started to examine the persuasion techniques used in conspiracy messages such as uncertainty and fear yet, little is understood about the visual narratives, especially how visual narratives differ in videos that debunk conspiracies versus those that propagate conspiracies. This paper addresses this gap in understanding visual framing in conspiracy videos through analyzing millions of frames from conspiracy and counter-conspiracy YouTube videos using computational methods. We found that conspiracy videos tended to use lower color variance and brightness, especially in thumbnails and earlier parts of the videos. This paper also demonstrates how researchers can integrate textual and visual features in machine learning models to study conspiracies on social media and discusses the implications of computational modeling for scholars interested in studying visual manipulation in the digital era. The analysis of visual and textual features presented in this paper could be useful for future studies focused on designing systems to identify conspiracy content on the Internet.

(0 min) The spinal cord (SC), which conveys information between the brain and the peripheral nervous system, plays a key role in various neurological disorders such as multiple sclerosis (MS) and amyotrophic lateral sclerosis (ALS), in which both gray matter (GM) and white matter (WM) may be impaired. While automated methods for WM/GM segmentation are now largely available, these techniques, developed for conventional systems (3T or lower) do not necessarily perform well on 7T MRI data, which feature finer details, contrasts, but also different artifacts or signal dropout. The primary goal of this study is thus to propose a new deep learning model that allows robust SC/GM multi-class segmentation based on ultra-high resolution 7T T2*-w MR images. The second objective is to highlight the relevance of implementing a specific data augmentation (DA) strategy, in particular to generate a generic model that could be used for multi-center studies at 7T.

(0 min) The conventional wisdom behind learning deep classification models is to focus on bad-classified examples and ignore well-classified examples that are far from the decision boundary. For instance, when training with cross-entropy loss, examples with higher likelihoods (i.e., well-classified examples) contribute smaller gradients in back-propagation. However, we theoretically show that this common practice hinders representation learning, energy optimization, and the growth of margin. To counteract this deficiency, we propose to reward well-classified examples with additive bonuses to revive their contribution to learning. This counterexample theoretically addresses these three issues. We empirically support this claim by directly verify the theoretical results or through the significant performance improvement with our counterexample on diverse tasks, including image classification, graph classification, and machine translation. Furthermore, this paper shows that because our idea can solve these three issues, we can deal with complex scenarios, such as imbalanced classification, OOD detection, and applications under adversarial attacks.

(0 min) Object detection is a basic and important task in the field of aerial image processing and has gained much attention in computer vision. However, previous aerial image object detection approaches have insufficient use of scene semantic information between different regions of large-scale aerial images. In addition, complex background and scale changes make it difficult to improve detection accuracy. To address these issues, we propose a relationship representation network for object detection in aerial images (RelationRS): 1) Firstly, multi-scale features are fused and enhanced by a dual relationship module (DRM) with conditional convolution. The dual relationship module learns the potential relationship between features of different scales and learns the relationship between different scenes from different patches in a same iteration. In addition, the dual relationship module dynamically generates parameters to guide the fusion of multi-scale features. 2) Secondly, The bridging visual representations module (BVR) is introduced into the field of aerial images to improve the object detection effect in images with complex backgrounds. Experiments with a publicly available object detection dataset for aerial images demonstrate that the proposed RelationRS achieves a state-of-the-art detection performance.

(0 min) The goal in a blind image quality assessment (BIQA) model is to simulate the process of evaluating images by human eyes and accurately assess the quality of the image. Although many approaches effectively identify degradation, they do not fully consider the semantic content in images resulting in distortion. In order to fill this gap, we propose a deep adaptive superpixel-based network, namely DSN-IQA, to assess the quality of image based on multi-scale and superpixel segmentation. The DSN-IQA can adaptively accept arbitrary scale images as input images, making the assessment process similar to human perception. The network uses two models to extract multi-scale semantic features and generate a superpixel adjacency map. These two elements are united together via feature fusion to accurately predict image quality. Experimental results on different benchmark databases demonstrate that our algorithm is highly competitive with other approaches when assessing challenging authentic image databases. Also, due to adaptive deep superpixel-based network, our model accurately assesses images with complicated distortion, much like the human eye.

(0 min) In this paper, we propose THOMAS, a joint multi-agent trajectory prediction framework allowing for efficient and consistent prediction of multi-agent multi-modal trajectories. We present a unified model architecture for fast and simultaneous agent future heatmap estimation leveraging hierarchical and sparse image generation. We demonstrate that heatmap output enables a higher level of control on the predicted trajectories compared to vanilla multi-modal trajectory regression, allowing to incorporate additional constraints for tighter sampling or collision-free predictions in a deterministic way. However, we also highlight that generating scene-consistent predictions goes beyond the mere generation of collision-free trajectories. We therefore propose a learnable trajectory recombination model that takes as input a set of predicted trajectories for each agent and outputs its consistent reordered recombination. We report our results on the Interaction multi-agent prediction challenge and rank $1^{st}$ on the online test leaderboard.

(0 min) Supervised Pix2Pix and unsupervised Cycle-consistency are two modes that dominate the field of medical image-to-image translation. However, neither modes are ideal. The Pix2Pix mode has excellent performance. But it requires paired and well pixel-wise aligned images, which may not always be achievable due to respiratory motion or anatomy change between times that paired images are acquired. The Cycle-consistency mode is less stringent with training data and works well on unpaired or misaligned images. But its performance may not be optimal. In order to break the dilemma of the existing modes, we propose a new unsupervised mode called RegGAN for medical image-to-image translation. It is based on the theory of "loss-correction". In RegGAN, the misaligned target images are considered as noisy labels and the generator is trained with an additional registration network to fit the misaligned noise distribution adaptively. The goal is to search for the common optimal solution to both image-to-image translation and registration tasks. We incorporated RegGAN into a few state-of-the-art image-to-image translation methods and demonstrated that RegGAN could be easily combined with these methods to improve their performances. Such as a simple CycleGAN in our mode surpasses latest NICEGAN even though using less network parameters. Based on our results, RegGAN outperformed both Pix2Pix on aligned data and Cycle-consistency on misaligned or unpaired data. RegGAN is insensitive to noises which makes it a better choice for a wide range of scenarios, especially for medical image-to-image translation tasks in which well pixel-wise aligned data are not available

(0 min) Due to the rapid increase in the diversity of image data, the problem of domain generalization has received increased attention recently. While domain generalization is a challenging problem, it has achieved great development thanks to the fast development of AI techniques in computer vision. Most of these advanced algorithms are proposed with deep architectures based on convolution neural nets (CNN). However, though CNNs have a strong ability to find the discriminative features, they do a poor job of modeling the relations between different locations in the image due to the response to CNN filters are mostly local. Since these local and global spatial relationships are characterized to distinguish an object under consideration, they play a critical role in improving the generalization ability against the domain gap. In order to get the object parts relationships to gain better domain generalization, this work proposes to use the self attention model. However, the attention models are proposed for sequence, which are not expert in discriminate feature extraction for 2D images. Considering this, we proposed a hybrid architecture to discover the spatial relationships between these local features, and derive a composite representation that encodes both the discriminative features and their relationships to improve the domain generalization. Evaluation on three well-known benchmarks demonstrates the benefits of modeling relationships between the features of an image using the proposed method and achieves state-of-the-art domain generalization performance. More specifically, the proposed algorithm outperforms the state-of-the-art by 2.2% and 3.4% on PACS and Office-Home databases, respectively.

(0 min) The new coronavirus (known as COVID-19) was first identified in Wuhan and quickly spread worldwide, wreaking havoc on the economy and people's everyday lives. Fever, cough, sore throat, headache, exhaustion, muscular aches, and difficulty breathing are all typical symptoms of COVID-19. A reliable detection technique is needed to identify affected individuals and care for them in the early stages of COVID-19 and reduce the virus's transmission. The most accessible method for COVID-19 identification is RT-PCR; however, due to its time commitment and false-negative results, alternative options must be sought. Indeed, compared to RT-PCR, chest CT scans and chest X-ray images provide superior results. Because of the scarcity and high cost of CT scan equipment, X-ray images are preferable for screening. In this paper, a pre-trained network, DenseNet169, was employed to extract features from X-ray images. Features were chosen by a feature selection method (ANOVA) to reduce computations and time complexity while overcoming the curse of dimensionality to improve predictive accuracy. Finally, selected features were classified by XGBoost. The ChestX-ray8 dataset, which was employed to train and evaluate the proposed method. This method reached 98.72% accuracy for two-class classification (COVID-19, healthy) and 92% accuracy for three-class classification (COVID-19, healthy, pneumonia).

(0 min) Coronavirus disease 2019 (COVID-19) is the highly contagious illness caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The standard diagnostic testing procedure for COVID-19 is testing a nasopharyngeal swab for SARS-CoV-2 nucleic acid using a real-time polymerase chain reaction (PCR), which can take multiple days to provide a diagnosis. Another widespread form of testing is rapid antigen testing, which has a low sensitivity compared to PCR, but is favored for its quick diagnosis time of usually 15-30 minutes. Patients who test positive for COVID-19 demonstrate diffuse alveolar damage in 87% of cases. Machine learning has proven to have advantages in image classification problems with radiology. In this work, we introduce CovXR as a machine learning model designed to detect COVID-19 pneumonia in chest X-rays (CXR). CovXR is a convolutional neural network (CNN) trained on over 4,300 chest X-rays. The performance of the model is measured through accuracy, F1 score, sensitivity, and specificity. The model achieves an accuracy of 95.5% and an F1 score of 0.954. The sensitivity is 93.5% and specificity is 97.5%. With accuracy above 95% and F1 score above 0.95, CovXR is highly accurate in predicting COVID-19 pneumonia on CXRs. The model achieves better accuracy than prior work and uses a unique approach to identify COVID-19 pneumonia. CovXR is highly accurate in identifying COVID-19 on CXRs of patients with a PCR confirmed positive diagnosis and provides much faster results than PCR tests.

(0 min) Although recent semantic segmentation methods have made remarkable progress, they still rely on large amounts of annotated training data, which are often infeasible to collect in the autonomous driving scenario. Previous works usually tackle this issue with Unsupervised Domain Adaptation (UDA), which entails training a network on synthetic images and applying the model to real ones while minimizing the discrepancy between the two domains. Yet, these techniques do not consider additional information that may be obtained from other tasks. Differently, we propose to exploit self-supervised monocular depth estimation to improve UDA for semantic segmentation. On one hand, we deploy depth to realize a plug-in component which can inject complementary geometric cues into any existing UDA method. We further rely on depth to generate a large and varied set of samples to Self-Train the final model. Our whole proposal allows for achieving state-of-the-art performance (58.8 mIoU) in the GTA5->CS benchmark benchmark. Code is available at https://github.com/CVLAB-Unibo/d4-dbst.

(2 min) The traditional recommendation systems mainly use offline user data to train offline models, and then recommend items for online users, thus suffering from the unreliable estimation of user preferences based on sparse and noisy historical data. Conversational Recommendation System (CRS) uses the interactive form of the dialogue systems to solve the intrinsic problems of traditional recommendation systems. However, due to the lack of contextual information modeling, the existing CRS models are unable to deal with the exploitation and exploration (E&E) problem well, resulting in the heavy burden on users. To address the aforementioned issue, this work proposes a contextual information enhancement model tailored for CRS, called Knowledge Graph-enhanced Sampling (KGenSam). KGenSam integrates the dynamic graph of user interaction data with the external knowledge into one heterogeneous Knowledge Graph (KG) as the contextual information environment. Then, two samplers are designed to enhance knowledge by sampling fuzzy samples with high uncertainty for obtaining user preferences and reliable negative samples for updating recommender to achieve efficient acquisition of user preferences and model updating, and thus provide a powerful solution for CRS to deal with E&E problem. Experimental results on two real-world datasets demonstrate the superiority of KGenSam with significant improvements over state-of-the-art methods.

(2 min) POI recommendation is a key task in tourism information systems. However, in contrast to conventional point of interest (POI) recommender systems, the available data is extremely sparse; most tourist visit a few sightseeing spots once and most of these spots have no check-in data from new tourists. Most conventional systems rank sightseeing spots based on their popularity, reputations, and category-based similarities with users' preferences. They do not clarify what users can experience in these spots, which makes it difficult to meet diverse tourism needs. To this end, in this work, we propose a mechanism to recommend POIs to tourists. Our mechanism include two components: one is a probabilistic model that reveals the user behaviors in tourism; the other is a pseudo rating mechanism to handle the cold-start issue in POIs recommendations. We carried out extensive experiments with two datasets collected from Flickr. The experimental results demonstrate that our methods are superior to the state-of-the-art methods in both the recommendation performances (precision, recall and F-measure) and fairness. The experimental results also validate the robustness of the proposed methods, i.e., our methods can handle well the issue of data sparsity.

(2 min) Growth in technology has resulted in the large-scale collection and processing of Personally Identifiable Information by organizations that run digital services such as websites, which led to the emergence of new legislation to regulate PII collection and processing by organizations. Subsequently, several African countries have recently started enacting new data protection regulations due to recent technological innovations. However, there is little information about the security and privacy practices of top websites serving content to EAC citizens. We, therefore, analyze the website operators' patterns in terms of third-party tracking, security of data transmission, cookie information, and privacy policies for 169 top EAC website operators using WebXray, OpenSSL, and Alexa top websites API. Our results show that only 75 percent of the analyzed websites have a privacy policy in place. Out of this, only 16 percent of the third-party tracking companies that track users on a particular website are disclosed in the site's privacy policy statements which means that users don not have a way of knowing which third parties collect data about them when they visit a website. Such privacy policies take time to read and are difficult to understand; on average, it takes a college graduate to comprehend the policy and a user spends 12 minutes to read the policy. Additionally, most third-party tracking on EAC websites is related to advertisement and belongs to companies outside the EAC. This means that EAC lawmakers need to enact suitable laws to ensure that people's privacy is protected as the rate of technology adoption continues to increase.

(2 min) Despite their recent popularity and well known advantages, dense retrievers still lag behind sparse methods such as BM25 in their ability to reliably match salient phrases and rare entities in the query. It has been argued that this is an inherent limitation of dense models. We disprove this claim by introducing the Salient Phrase Aware Retriever (SPAR), a dense retriever with the lexical matching capacity of a sparse model. In particular, we show that a dense retriever {\Lambda} can be trained to imitate a sparse one, and SPAR is built by augmenting a standard dense retriever with {\Lambda}. When evaluated on five open-domain question answering datasets and the MS MARCO passage retrieval task, SPAR sets a new state of the art for dense and sparse retrievers and can match or exceed the performance of more complicated dense-sparse hybrid systems.

(2 min) The evaluation of question answering models compares ground-truth annotations with model predictions. However, as of today, this comparison is mostly lexical-based and therefore misses out on answers that have no lexical overlap but are still semantically similar, thus treating correct answers as false. This underestimation of the true performance of models hinders user acceptance in applications and complicates a fair comparison of different models. Therefore, there is a need for an evaluation metric that is based on semantics instead of pure string similarity. In this short paper, we present SAS, a cross-encoder-based metric for the estimation of semantic answer similarity, and compare it to seven existing metrics. To this end, we create an English and a German three-way annotated evaluation dataset containing pairs of answers along with human judgment of their semantic similarity, which we release along with an implementation of the SAS metric and the experiments. We find that semantic similarity metrics based on recent transformer models correlate much better with human judgment than traditional lexical similarity metrics on our two newly created datasets and one dataset from related work.

(2 min) Personalized outfit recommendation has recently been in the spotlight with the rapid growth of the online fashion industry. However, recommending outfits has two significant challenges that should be addressed. The first challenge is that outfit recommendation often requires a complex and large model that utilizes visual information, incurring huge memory and time costs. One natural way to mitigate this problem is to compress such a cumbersome model with knowledge distillation (KD) techniques that leverage knowledge from a pretrained teacher model. However, it is hard to apply existing KD approaches in recommender systems (RS) to the outfit recommendation because they require the ranking of all possible outfits while the number of outfits grows exponentially to the number of consisting clothing items. Therefore, we propose a new KD framework for outfit recommendation, called False Negative Distillation (FND), which exploits false-negative information from the teacher model while not requiring the ranking of all candidates. The second challenge is that the explosive number of outfit candidates amplifying the data sparsity problem, often leading to poor outfit representation. To tackle this issue, inspired by the recent success of contrastive learning (CL), we introduce a CL framework for outfit representation learning with two proposed data augmentation methods. Quantitative and qualitative experiments on outfit recommendation datasets demonstrate the effectiveness and soundness of our proposed methods.

(2 min) We propose a novel method for applying Transformer models to extractive question answering (QA) tasks. Recently, pretrained generative sequence-to-sequence (seq2seq) models have achieved great success in question answering. Contributing to the success of these models are internal attention mechanisms such as cross-attention. We propose a simple strategy to obtain an extractive answer span from the generative model by leveraging the decoder cross-attention patterns. Viewing cross-attention as an architectural prior, we apply joint training to further improve QA performance. Empirical results show that on open-domain question answering datasets like NaturalQuestions and TriviaQA, our method approaches state-of-the-art performance on both generative and extractive inference, all while using much fewer parameters. Furthermore, this strategy allows us to perform hallucination-free inference while conferring significant improvements to the model's ability to rerank relevant passages.

(2 min) Recent years have witnessed the dramatic growth of paper volumes with plenty of new research papers published every day, especially in the area of computer science. How to glean papers worth reading from the massive literature to do a quick survey or keep up with the latest advancement about a specific research topic has become a challenging task. Existing academic search engines such as Google Scholar return relevant papers by individually calculating the relevance between each paper and query. However, such systems usually omit the prerequisite chains of a research topic and cannot form a meaningful reading path. In this paper, we introduce a new task named Reading Path Generation (RPG) which aims at automatically producing a path of papers to read for a given query. To serve as a research benchmark, we further propose SurveyBank, a dataset consisting of large quantities of survey papers in the field of computer science as well as their citation relationships. Each survey paper contains key phrases extracted from its title and multi-level reading lists inferred from its references. Furthermore, we propose a graph-optimization-based approach for reading path generation which takes the relationship between papers into account. Extensive evaluations demonstrate that our approach outperforms other baselines. A Real-time Reading Path Generation System (RePaGer) has been also implemented with our designed model. To the best of our knowledge, we are the first to target this important research problem. Our source code of RePaGer system and SurveyBank dataset can be found on here.

(2 min) The travel marketing platform of Alibaba serves an indispensable role for hundreds of different travel scenarios from Fliggy, Taobao, Alipay apps, etc. To provide personalized recommendation service for users visiting different scenarios, there are two critical issues to be carefully addressed. First, since the traffic characteristics of different scenarios, it is very challenging to train a unified model to serve all. Second, during the promotion period, the exposure of some specific items will be re-weighted due to manual intervention, resulting in biased logs, which will degrade the ranking model trained using these biased data. In this paper, we propose a novel Scenario-Aware Ranking Network (SAR-Net) to address these issues. SAR-Net harvests the abundant data from different scenarios by learning users' cross-scenario interests via two specific attention modules, which leverage the scenario features and item features to modulate the user behavior features, respectively. Then, taking the encoded features of previous module as input, a scenario-specific linear transformation layer is adopted to further extract scenario-specific features, followed by two groups of debias expert networks, i.e., scenario-specific experts and scenario-shared experts. They output intermediate results independently, which are further fused into the final result by a multi-scenario gating module. In addition, to mitigate the data fairness issue caused by manual intervention, we propose the concept of Fairness Coefficient (FC) to measures the importance of individual sample and use it to reweigh the prediction in the debias expert networks. Experiments on an offline dataset covering over 80 million users and 1.55 million travel items and an online A/B test demonstrate the effectiveness of our SAR-Net and its superiority over state-of-the-art methods.

(2 min) The key to personalized news recommendation is to match the user's interests with the candidate news precisely and efficiently. Most existing approaches embed user interests into a representation vector then recommend by comparing it with the candidate news vector. In such a workflow, fine-grained matching signals may be lost. Recent studies try to cover that by modeling fine-grained interactions between the candidate news and each browsed news article of the user. Despite the effectiveness improvement, these models suffer from much higher computation costs online. Consequently, it remains a tough issue to take advantage of effective interactions in an efficient way. To address this problem, we proposed an end-to-end Selective Fine-grained Interaction framework (SFI) with a learning-to-select mechanism. Instead of feeding all historical news into interaction, SFI can quickly select informative historical news w.r.t. the candidate and exclude others from following computations. We empower the selection to be both sparse and automatic, which guarantees efficiency and effectiveness respectively. Extensive experiments on the publicly available dataset MIND validates the superiority of SFI over the state-of-the-art methods: with only five historical news selected, it can significantly improve the AUC by 2.17% over the state-of-the-art interaction-based models; at the same time, it is four times faster.

(2 min) To capture the semantic graph structure from raw text, most existing summarization approaches are built on GNNs with a pre-trained model. However, these methods suffer from cumbersome procedures and inefficient computations for long-text documents. To mitigate these issues, this paper proposes HETFORMER, a Transformer-based pre-trained model with multi-granularity sparse attentions for long-text extractive summarization. Specifically, we model different types of semantic nodes in raw text as a potential heterogeneous graph and directly learn heterogeneous relationships (edges) among nodes by Transformer. Extensive experiments on both single- and multi-document summarization tasks show that HETFORMER achieves state-of-the-art performance in Rouge F1 while using less memory and fewer parameters.

(2 min) In this paper, we propose a data-driven leak localization method for water distribution networks (WDNs) which combines two complementary approaches: graph-based interpolation and dictionary classification. The former estimates the complete WDN hydraulic state (i.e., hydraulic heads) from real measurements at certain nodes and the network graph. Then, these actual measurements, together with a subset of valuable estimated states, are used to feed and train the dictionary learning scheme. Thus, the meshing of these two methods is explored, showing that its performance is superior to either approach alone, even deriving different mechanisms to increase its resilience to classical problems (e.g., dimensionality, interpolation errors, etc.). The approach is validated using the L-TOWN benchmark proposed at BattLeDIM2020.

(2 min) This paper introduces S3PRL-VC, an open-source voice conversion (VC) framework based on the S3PRL toolkit. In the context of recognition-synthesis VC, self-supervised speech representation (S3R) is valuable in its potential to replace the expensive supervised representation adopted by state-of-the-art VC systems. Moreover, we claim that VC is a good probing task for S3R analysis. In this work, we provide a series of in-depth analyses by benchmarking on the two tasks in VCC2020, namely intra-/cross-lingual any-to-one (A2O) VC, as well as an any-to-any (A2A) setting. We also provide comparisons between not only different S3Rs but also top systems in VCC2020 with supervised representations. Systematic objective and subjective evaluation were conducted, and we show that S3R is comparable with VCC2020 top systems in the A2O setting in terms of similarity, and achieves state-of-the-art in S3R-based A2A VC. We believe the extensive analysis, as well as the toolkit itself, contribute to not only the S3R community but also the VC community. The codebase is now open-sourced.

(2 min) We consider the problem of uniformity testing of Lipschitz continuous distributions with bounded support. The alternative hypothesis is a composite set of Lipschitz continuous distributions that are at least $\varepsilon$ away in $\ell_1$ distance from the uniform distribution. We propose a sequential test that adapts to the unknown distribution under the alternative hypothesis. Referred to as the Adaptive Binning Coincidence (ABC) test, the proposed strategy adapts in two ways. First, it partitions the set of alternative distributions into layers based on their distances to the uniform distribution. It then sequentially eliminates the alternative distributions layer by layer in decreasing distance to the uniform, and subsequently takes advantage of favorable situations of a distant alternative by exiting early. Second, it adapts, across layers of the alternative distributions, the resolution level of the discretization for computing the coincidence statistic. The farther away the layer is from the uniform, the coarser the discretization is needed for eliminating/exiting this layer. It thus exits both early in the detection process and quickly by using a lower resolution to take advantage of favorable alternative distributions. The ABC test builds on a novel sequential coincidence test for discrete distributions, which is of independent interest. We establish the sample complexity of the proposed tests as well as a lower bound.

(2 min) Consistency training is a popular method to improve deep learning models in computer vision and natural language processing. Graph neural networks (GNNs) have achieved remarkable performance in a variety of network science learning tasks, but to date no work has studied the effect of consistency training on large-scale graph problems. GNNs scale to large graphs by minibatch training and subsample node neighbors to deal with high degree nodes. We utilize the randomness inherent in the subsampling of neighbors and introduce a novel consistency training method to improve accuracy. For a target node we generate different neighborhood expansions, and distill the knowledge of the average of the predictions to the GNN. Our method approximates the expected prediction of the possible neighborhood samples and practically only requires a few samples. We demonstrate that our training method outperforms standard GNN training in several different settings, and yields the largest gains when label rates are low.

(2 min) In multimodal assistant, where vision is also one of the input modalities, the identification of user intent becomes a challenging task as visual input can influence the outcome. Current digital assistants take spoken input and try to determine the user intent from conversational or device context. So, a dataset, which includes visual input (i.e. images or videos for the corresponding questions targeted for multimodal assistant use cases, is not readily available. The research in visual question answering (VQA) and visual question generation (VQG) is a great step forward. However, they do not capture questions that a visually-abled person would ask multimodal assistants. Moreover, many times questions do not seek information from external knowledge. In this paper, we provide a new dataset, MMIU (MultiModal Intent Understanding), that contains questions and corresponding intents provided by human annotators while looking at images. We, then, use this dataset for intent classification task in multimodal digital assistant. We also experiment with various approaches for combining vision and language features including the use of multimodal transformer for classification of image-question pairs into 14 intents. We provide the benchmark results and discuss the role of visual and text features for the intent classification task on our dataset.

(2 min) Ultrasound imaging is a commonly used technology for visualising patient anatomy in real-time during diagnostic and therapeutic procedures. High operator dependency and low reproducibility make ultrasound imaging and interpretation challenging with a steep learning curve. Automatic image classification using deep learning has the potential to overcome some of these challenges by supporting ultrasound training in novices, as well as aiding ultrasound image interpretation in patient with complex pathology for more experienced practitioners. However, the use of deep learning methods requires a large amount of data in order to provide accurate results. Labelling large ultrasound datasets is a challenging task because labels are retrospectively assigned to 2D images without the 3D spatial context available in vivo or that would be inferred while visually tracking structures between frames during the procedure. In this work, we propose a multi-modal convolutional neural network (CNN) architecture that labels endoscopic ultrasound (EUS) images from raw verbal comments provided by a clinician during the procedure. We use a CNN composed of two branches, one for voice data and another for image data, which are joined to predict image labels from the spoken names of anatomical landmarks. The network was trained using recorded verbal comments from expert operators. Our results show a prediction accuracy of 76% at image level on a dataset with 5 different labels. We conclude that the addition of spoken commentaries can increase the performance of ultrasound image classification, and eliminate the burden of manually labelling large EUS datasets necessary for deep learning applications.

(2 min) Conventional supervised learning typically assumes that the learning task can be solved by learning a single function since the data is sampled from a fixed distribution. However, this assumption is invalid in open-ended environments where no task-level data partitioning is available. In this paper, we present a novel supervised learning framework of learning from open-ended data, which is modeled as data implicitly sampled from multiple domains with the data in each domain obeying a domain-specific target function. Since different domains may possess distinct target functions, open-ended data inherently requires multiple functions to capture all its input-output relations, rendering training a single global model problematic. To address this issue, we devise an Open-ended Supervised Learning (OSL) framework, of which the key component is a subjective function that allocates the data among multiple candidate models to resolve the "conflict" between the data from different domains, exhibiting a natural hierarchy. We theoretically analyze the learnability and the generalization error of OSL, and empirically validate its efficacy in both open-ended regression and classification tasks.

(2 min) Randomized Smoothing (RS) is a promising method for obtaining robustness certificates by evaluating a base model under noise. In this work, we: (i) theoretically motivate why ensembles are a particularly suitable choice as base models for RS, and (ii) empirically confirm this choice, obtaining state-of-the-art results in multiple settings. The key insight of our work is that the reduced variance of ensembles over the perturbations introduced in RS leads to significantly more consistent classifications for a given input. This, in turn, leads to substantially increased certifiable radii for samples close to the decision boundary. Additionally, we introduce key optimizations which enable an up to 55-fold decrease in sample complexity of RS, thus drastically reducing its computational overhead. Experimentally, we show that ensembles of only 3 to 10 classifiers consistently improve on their strongest constituting model with respect to their average certified radius (ACR) by 5% to 21% on both CIFAR10 and ImageNet, achieving a new state-of-the-art ACR of 0.86 and 1.11, respectively. We release all code and models required to reproduce our results upon publication.

(2 min) Modern reinforcement learning (RL) commonly engages practical problems with large state spaces, where function approximation must be deployed to approximate either the value function or the policy. While recent progresses in RL theory address a rich set of RL problems with general function approximation, such successes are mostly restricted to the single-agent setting. It remains elusive how to extend these results to multi-agent RL, especially due to the new challenges arising from its game-theoretical nature. This paper considers two-player zero-sum Markov Games (MGs). We propose a new algorithm that can provably find the Nash equilibrium policy using a polynomial number of samples, for any MG with low multi-agent Bellman-Eluder dimension -- a new complexity measure adapted from its single-agent version (Jin et al., 2021). A key component of our new algorithm is the exploiter, which facilitates the learning of the main player by deliberately exploiting her weakness. Our theoretical framework is generic, which applies to a wide range of models including but not limited to tabular MGs, MGs with linear or kernel function approximation, and MGs with rich observations.

(2 min) Causal discovery, i.e., inferring underlying cause-effect relationships from observations of a scene or system, is an inherent mechanism in human cognition, but has been shown to be highly challenging to automate. The majority of approaches in the literature aiming for this task consider constrained scenarios with fully observed variables or data from stationary time-series. In this work we aim for causal discovery in a more general class of scenarios, scenes with non-stationary behavior over time. For our purposes we here regard a scene as a composition objects interacting with each other over time. Non-stationarity is modeled as stationarity conditioned on an underlying variable, a state, which can be of varying dimension, more or less hidden given observations of the scene, and also depend more or less directly on these observations. We propose a probabilistic deep learning approach called State-Dependent Causal Inference (SDCI) for causal discovery in such conditionally stationary time-series data. Results in two different synthetic scenarios show that this method is able to recover the underlying causal dependencies with high accuracy even in cases with hidden states.

(2 min) Numerous recent works utilize bi-Lipschitz regularization of neural network layers to preserve relative distances between data instances in the feature spaces of each layer. This distance sensitivity with respect to the data aids in tasks such as uncertainty calibration and out-of-distribution (OOD) detection. In previous works, features extracted with a distance sensitive model are used to construct feature covariance matrices which are used in deterministic uncertainty estimation or OOD detection. However, in cases where there is a distribution over tasks, these methods result in covariances which are sub-optimal, as they may not leverage all of the meta information which can be shared among tasks. With the use of an attentive set encoder, we propose to meta learn either diagonal or diagonal plus low-rank factors to efficiently construct task specific covariance matrices. Additionally, we propose an inference procedure which utilizes scaled energy to achieve a final predictive distribution which can better separate OOD data, and is well calibrated under a distributional dataset shift.

(2 min) In most scientific domains, the deep learning community has largely focused on the quality of deep generative models, resulting in highly accurate and successful solutions. However, this race for quality comes at a tremendous computational cost, which incurs vast energy consumption and greenhouse gas emissions. At the heart of this problem are the measures that we use as a scientific community to evaluate our work. In this paper, we suggest relying on a multi-objective measure based on Pareto optimality, which takes into account both the quality of the model and its energy consumption. By applying our measure on the current state-of-the-art in generative audio models, we show that it can drastically change the significance of the results. We believe that this type of metric can be widely used by the community to evaluate their work, while putting computational cost -- and in fine energy consumption -- in the spotlight of deep learning research.

(2 min) This report describes our solution to the VALUE Challenge 2021 in the captioning task. Our solution, named CLIP4Caption++, is built on X-Linear/X-Transformer, which is an advanced model with encoder-decoder architecture. We make the following improvements on the proposed CLIP4Caption++: We employ an advanced encoder-decoder model architecture X-Transformer as our main framework and make the following improvements: 1) we utilize three strong pre-trained CLIP models to extract the text-related appearance visual features. 2) we adopt the TSN sampling strategy for data enhancement. 3) we involve the video subtitle information to provide richer semantic information. 3) we introduce the subtitle information, which fuses with the visual features as guidance. 4) we design word-level and sentence-level ensemble strategies. Our proposed method achieves 86.5, 148.4, 64.5 CIDEr scores on VATEX, YC2C, and TVC datasets, respectively, which shows the superior performance of our proposed CLIP4Caption++ on all three datasets.

(2 min) Motivation: MicroRNAs (miRNAs) play pivotal roles in gene expression regulation by binding to target sites of messenger RNAs (mRNAs). While identifying functional targets of miRNAs is of utmost importance, their prediction remains a great challenge. Previous computational algorithms have major limitations. They use conservative candidate target site (CTS) selection criteria mainly focusing on canonical site types, rely on laborious and time-consuming manual feature extraction, and do not fully capitalize on the information underlying miRNA-CTS interactions. Results: In this paper, we introduce TargetNet, a novel deep learning-based algorithm for functional miRNA target prediction. To address the limitations of previous approaches, TargetNet has three key components: (1) relaxed CTS selection criteria accommodating irregularities in the seed region, (2) a novel miRNA-CTS sequence encoding scheme incorporating extended seed region alignments, and (3) a deep residual network-based prediction model. The proposed model was trained with miRNA-CTS pair datasets and evaluated with miRNA-mRNA pair datasets. TargetNet advances the previous state-of-the-art algorithms used in functional miRNA target classification. Furthermore, it demonstrates great potential for distinguishing high-functional miRNA targets.

(2 min) Community detection is an important research topic in graph analytics that has a wide range of applications. A variety of static community detection algorithms and quality metrics were developed in the past few years. However, most real-world graphs are not static and often change over time. In the case of streaming data, communities in the associated graph need to be updated either continuously or whenever new data streams are added to the graph, which poses a much greater challenge in devising good community detection algorithms for maintaining dynamic graphs over streaming data. In this paper, we propose an incremental community detection algorithm for maintaining a dynamic graph over streaming data. The contributions of this study include (a) the implementation of a Distributed Weighted Community Clustering (DWCC) algorithm, (b) the design and implementation of a novel Incremental Distributed Weighted Community Clustering (IDWCC) algorithm, and (c) an experimental study to compare the performance of our IDWCC algorithm with the DWCC algorithm. We validate the functionality and efficiency of our framework in processing streaming data and performing large in-memory distributed dynamic graph analytics. The results demonstrate that our IDWCC algorithm performs up to three times faster than the DWCC algorithm for a similar accuracy.

(2 min) Optimal Transport (OT) metrics allow for defining discrepancies between two probability measures. Wasserstein distance is for longer the celebrated OT-distance frequently-used in the literature, which seeks probability distributions to be supported on the $\textit{same}$ metric space. Because of its high computational complexity, several approximate Wasserstein distances have been proposed based on entropy regularization or on slicing, and one-dimensional Wassserstein computation. In this paper, we propose a novel extension of Wasserstein distance to compare two incomparable distributions, that hinges on the idea of $\textit{distributional slicing}$, embeddings, and on computing the closed-form Wassertein distance between the sliced distributions. We provide a theoretical analysis of this new divergence, called $\textit{heterogeneous Wasserstein discrepancy (HWD)}$, and we show that it preserves several interesting properties including rotation-invariance. We show that the embeddings involved in HWD can be efficiently learned. Finally, we provide a large set of experiments illustrating the behavior of HWD as a divergence in the context of generative modeling and in query framework.

(2 min) For certain infinitely-wide neural networks, the neural tangent kernel (NTK) theory fully characterizes generalization, but for the networks used in practice, the empirical NTK only provides a rough first-order approximation. Still, a growing body of work keeps leveraging this approximation to successfully analyze important deep learning phenomena and design algorithms for new applications. In our work, we provide strong empirical evidence to determine the practical validity of such approximation by conducting a systematic comparison of the behavior of different neural networks and their linear approximations on different tasks. We show that the linear approximations can indeed rank the learning complexity of certain tasks for neural networks, even when they achieve very different performances. However, in contrast to what was previously reported, we discover that neural networks do not always perform better than their kernel approximations, and reveal that the performance gap heavily depends on architecture, dataset size and training task. We discover that networks overfit to these tasks mostly due to the evolution of their kernel during training, thus, revealing a new type of implicit bias.

(2 min) Understanding a decision-maker's priorities by observing their behavior is critical for transparency and accountability in decision processes, such as in healthcare. Though conventional approaches to policy learning almost invariably assume stationarity in behavior, this is hardly true in practice: Medical practice is constantly evolving as clinical professionals fine-tune their knowledge over time. For instance, as the medical community's understanding of organ transplantations has progressed over the years, a pertinent question is: How have actual organ allocation policies been evolving? To give an answer, we desire a policy learning method that provides interpretable representations of decision-making, in particular capturing an agent's non-stationary knowledge of the world, as well as operating in an offline manner. First, we model the evolving behavior of decision-makers in terms of contextual bandits, and formalize the problem of Inverse Contextual Bandits ("ICB"). Second, we propose two concrete algorithms as solutions, learning parametric and nonparametric representations of an agent's behavior. Finally, using both real and simulated data for liver transplantations, we illustrate the applicability and explainability of our method, as well as benchmarking and validating the accuracy of our algorithms.

(2 min) Finding a suitable density function is essential for density-based clustering algorithms such as DBSCAN and DPC. A naive density corresponding to the indicator function of a unit $d$-dimensional Euclidean ball is commonly used in these algorithms. Such density suffers from capturing local features in complex datasets. To tackle this issue, we propose a new kernel diffusion density function, which is adaptive to data of varying local distributional characteristics and smoothness. Furthermore, we develop a surrogate that can be efficiently computed in linear time and space and prove that it is asymptotically equivalent to the kernel diffusion density function. Extensive empirical experiments on benchmark and large-scale face image datasets show that the proposed approach not only achieves a significant improvement over classic density-based clustering algorithms but also outperforms the state-of-the-art face clustering methods by a large margin.

(2 min) Practical machine learning applications involving time series data, such as firewall log analysis to proactively detect anomalous behavior, are concerned with real time analysis of streaming data. Consequently, we need to update the ML models as the statistical characteristics of such data may shift frequently with time. One alternative explored in the literature is to retrain models with updated data whenever the models accuracy is observed to degrade. However, these methods rely on near real time availability of ground truth, which is rarely fulfilled. Further, in applications with seasonal data, temporal concept drift is confounded by seasonal variation. In this work, we propose an approach called Unsupervised Temporal Drift Detector or UTDD to flexibly account for seasonal variation, efficiently detect temporal concept drift in time series data in the absence of ground truth, and subsequently adapt our ML models to concept drift for better generalization.

(2 min) When learning tasks over time, artificial neural networks suffer from a problem known as Catastrophic Forgetting (CF). This happens when the weights of a network are overwritten during the training of a new task causing forgetting of old information. To address this issue, we propose MetA Reusable Knowledge or MARK, a new method that fosters weight reusability instead of overwriting when learning a new task. Specifically, MARK keeps a set of shared weights among tasks. We envision these shared weights as a common Knowledge Base (KB) that is not only used to learn new tasks, but also enriched with new knowledge as the model learns new tasks. Key components behind MARK are two-fold. On the one hand, a metalearning approach provides the key mechanism to incrementally enrich the KB with new knowledge and to foster weight reusability among tasks. On the other hand, a set of trainable masks provides the key mechanism to selectively choose from the KB relevant weights to solve each task. By using MARK, we achieve state of the art results in several popular benchmarks, surpassing the best performing methods in terms of average accuracy by over 10% on the 20-Split-MiniImageNet dataset, while achieving almost zero forgetfulness using 55% of the number of parameters. Furthermore, an ablation study provides evidence that, indeed, MARK is learning reusable knowledge that is selectively used by each task.

(2 min) Autoencoders are commonly used in representation learning. They consist of an encoder and a decoder, which provide a straightforward way to map n-dimensional data in input space to a lower m-dimensional representation space and back. The decoder itself defines an m-dimensional manifold in input space. Inspired by manifold learning, we show that the decoder can be trained on its own by learning the representations of the training samples along with the decoder weights using gradient descent. A sum-of-squares loss then corresponds to optimizing the manifold to have the smallest Euclidean distance to the training samples, and similarly for other loss functions. We derive expressions for the number of samples needed to specify the encoder and decoder and show that the decoder generally requires much less training samples to be well-specified compared to the encoder. We discuss training of autoencoders in this perspective and relate to previous work in the field that use noisy training examples and other types of regularization. On the natural image data sets MNIST and CIFAR10, we demonstrate that the decoder is much better suited to learn a low-dimensional representation, especially when trained on small data sets. Using simulated gene regulatory data, we further show that the decoder alone leads to better generalization and meaningful representations. Our approach of training the decoder alone facilitates representation learning even on small data sets and can lead to improved training of autoencoders. We hope that the simple analyses presented will also contribute to an improved conceptual understanding of representation learning.

(2 min) This paper introduces semi-ring dictionaries, a powerful class of compositional and purely functional collections that subsume other collection types such as sets, multisets, arrays, vectors, and matrices. We developed SDQL, a statically typed language that can express relational algebra with aggregations, linear algebra, and functional collections over data such as relations and matrices using semi-ring dictionaries. Furthermore, thanks to the algebraic structure behind these dictionaries, SDQL unifies a wide range of optimizations commonly used in databases (DB) and linear algebra (LA). As a result, SDQL enables efficient processing of hybrid DB and LA workloads, by putting together optimizations that are otherwise confined to either DB systems or LA frameworks. We show experimentally that a handful of DB and LA workloads can take advantage of the SDQL language and optimizations. Overall, we observe that SDQL achieves competitive performance relative to Typer and Tectorwise, which are state-of-the-art in-memory DB systems for (flat, not nested) relational data, and achieves an average 2x speedup over SciPy for LA workloads. For hybrid workloads involving LA processing, SDQL achieves up to one order of magnitude speedup over Trance, a state-of-the-art nested relational engine for nested biomedical data, and gives an average 40% speedup over LMFAO, a state-of-the-art in-DB machine learning engine for two (flat) relational real-world retail datasets.

(2 min) Oxygen consumption (VO$_2$) provides established clinical and physiological indicators of cardiorespiratory function and exercise capacity. However, VO$_2$ monitoring is largely limited to specialized laboratory settings, making its widespread monitoring elusive. Here, we investigate temporal prediction of VO$_2$ from wearable sensors during cycle ergometer exercise using a temporal convolutional network (TCN). Cardiorespiratory signals were acquired from a smart shirt with integrated textile sensors alongside ground-truth VO$_2$ from a metabolic system on twenty-two young healthy adults. Participants performed one ramp-incremental and three pseudorandom binary sequence exercise protocols to assess a range of VO$_2$ dynamics. A TCN model was developed using causal convolutions across an effective history length to model the time-dependent nature of VO$_2$. Optimal history length was determined through minimum validation loss across hyperparameter values. The best performing model encoded 218 s history length (TCN-VO$_2$ A), with 187 s, 97 s, and 76 s yielding less than 3% deviation from the optimal validation loss. TCN-VO$_2$ A showed strong prediction accuracy (mean, 95% CI) across all exercise intensities (-22 ml.min$^{-1}$, [-262, 218]), spanning transitions from low-moderate (-23 ml.min$^{-1}$, [-250, 204]), low-high (14 ml.min$^{-1}$, [-252, 280]), ventilatory threshold-high (-49 ml.min$^{-1}$, [-274, 176]), and maximal (-32 ml.min$^{-1}$, [-261, 197]) exercise. Second-by-second classification of physical activity across 16090 s of predicted VO$_2$ was able to discern between vigorous, moderate, and light activity with high accuracy (94.1%). This system enables quantitative aerobic activity monitoring in non-laboratory settings across a range of exercise intensities using wearable sensors for monitoring exercise prescription adherence and personal fitness.

(2 min) Inspired by its success in natural language processing and computer vision, pre-training has attracted substantial attention in cheminformatics and bioinformatics, especially for molecule based tasks. A molecule can be represented by either a graph (where atoms are connected by bonds) or a SMILES sequence (where depth-first-search is applied to the molecular graph with specific rules). Existing works on molecule pre-training use either graph representations only or SMILES representations only. In this work, we propose to leverage both the representations and design a new pre-training algorithm, dual-view molecule pre-training (briefly, DMP), that can effectively combine the strengths of both types of molecule representations. The model of DMP consists of two branches: a Transformer branch that takes the SMILES sequence of a molecule as input, and a GNN branch that takes a molecular graph as input. The training of DMP contains three tasks: (1) predicting masked tokens in a SMILES sequence by the Transformer branch, (2) predicting masked atoms in a molecular graph by the GNN branch, and (3) maximizing the consistency between the two high-level representations output by the Transformer and GNN branches separately. After pre-training, we can use either the Transformer branch (this one is recommended according to empirical results), the GNN branch, or both for downstream tasks. DMP is tested on nine molecular property prediction tasks and achieves state-of-the-art performances on seven of them. Furthermore, we test DMP on three retrosynthesis tasks and achieve state-of-the-art results on them.

(0 min) Deep neural networks tend to underestimate uncertainty and produce overly confident predictions. Recently proposed solutions, such as MC Dropout and SDENet, require complex training and/or auxiliary out-of-distribution data. We propose a simple solution by extending the time-tested iterative reweighted least square (IRLS) in generalised linear regression. We use two sub-networks to parametrise the prediction and uncertainty estimation, enabling easy handling of complex inputs and nonlinear response. The two sub-networks have shared representations and are trained via two complementary loss functions for the prediction and the uncertainty estimates, with interleaving steps as in a cooperative game. Compared with more complex models such as MC-Dropout or SDE-Net, our proposed network is simpler to implement and more robust (insensitive to varying aleatoric and epistemic uncertainty).

(2 min) Designing complex architectures has been an essential cogwheel in the revolution deep learning has brought about in the past decade. When solving difficult problems in a datadriven manner, a well-tried approach is to take an architecture discovered by renowned deep learning scientists as a basis (e.g. Inception) and try to apply it to a specific problem. This might be sufficient, but as of now, achieving very high accuracy on a complex or yet unsolved task requires the knowledge of highly-trained deep learning experts. In this work, we would like to contribute to the area of Automated Machine Learning (AutoML), specifically Neural Architecture Search (NAS), which intends to make deep learning methods available for a wider range of society by designing neural topologies automatically. Although several different approaches exist (e.g. gradient-based or evolutionary algorithms), our focus is on one of the most promising research directions, reinforcement learning. In this scenario, a recurrent neural network (controller) is trained to create problem-specific neural network architectures (child). The validation accuracies of the child networks serve as a reward signal for training the controller with reinforcement learning. The basis of our proposed work is Efficient Neural Architecture Search (ENAS), where parameter sharing is applied among the child networks. ENAS, like many other RL-based algorithms, emphasize the learning of child networks as increasing their convergence result in a denser reward signal for the controller, therefore significantly reducing training times. The controller was originally trained with REINFORCE. In our research, we propose to modify this to a more modern and complex algorithm, PPO, which has demonstrated to be faster and more stable in other environments. Then, we briefly discuss and evaluate our results.

(0 min) Optimal transport distances are powerful tools to compare probability distributions and have found many applications in machine learning. Yet their algorithmic complexity prevents their direct use on large scale datasets. To overcome this challenge, practitioners compute these distances on minibatches {\em i.e.} they average the outcome of several smaller optimal transport problems. We propose in this paper an analysis of this practice, which effects are not well understood so far. We notably argue that it is equivalent to an implicit regularization of the original problem, with appealing properties such as unbiased estimators, gradients and a concentration bound around the expectation, but also with defects such as loss of distance property. Along with this theoretical analysis, we also conduct empirical experiments on gradient flows, GANs or color transfer that highlight the practical interest of this strategy.

(0 min) Gaussian Processes (GPs) provide powerful probabilistic frameworks for interpolation, forecasting, and smoothing, but have been hampered by computational scaling issues. Here we investigate data sampled on one dimension (e.g., a scalar or vector time series sampled at arbitrarily-spaced intervals), for which state-space models are popular due to their linearly-scaling computational costs. It has long been conjectured that state-space models are general, able to approximate any one-dimensional GP. We provide the first general proof of this conjecture, showing that any stationary GP on one dimension with vector-valued observations governed by a Lebesgue-integrable continuous kernel can be approximated to any desired precision using a specifically-chosen state-space model: the Latent Exponentially Generated (LEG) family. This new family offers several advantages compared to the general state-space model: it is always stable (no unbounded growth), the covariance can be computed in closed form, and its parameter space is unconstrained (allowing straightforward estimation via gradient descent). The theorem's proof also draws connections to Spectral Mixture Kernels, providing insight about this popular family of kernels. We develop parallelized algorithms for performing inference and learning in the LEG model, test the algorithm on real and synthetic data, and demonstrate scaling to datasets with billions of samples.

(0 min) Software language models have achieved promising results predicting code completion usages, and several industry studies have described successful IDE integrations. Recently, accuracy in autocompletion prediction improved 12.8% from training on a real-world dataset collected from programmers' IDE activity. But what if limited examples of IDE autocompletion in the target programming language are available for model training? In this paper, we investigate the efficacy of pretraining autocompletion models on non-IDE, non-autocompletion, and different-language example code sequences. We find that these unsupervised pretrainings improve model accuracy by over 50% on very small fine-tuning datasets and over 10% on 50k labeled examples. We confirm the real-world impact of these pretrainings in an online setting through A/B testing on thousands of IDE autocompletion users, finding that pretraining is responsible for increases of up to 6.63% autocompletion usage.

(0 min) Through solving pretext tasks, self-supervised learning leverages unlabeled data to extract useful latent representations replacing traditional input features in the downstream task. In audio/speech signal processing, a wide range of features where engineered through decades of research efforts. As it turns out, learning to predict such features (a.k.a pseudo-labels) has proven to be a particularly relevant pretext task, leading to useful self-supervised representations which prove to be effective for downstream tasks. However, methods and common practices for combining such pretext tasks for better performance on the downstream task have not been explored and understood properly. In fact, the process relies almost exclusively on a computationally heavy experimental procedure, which becomes intractable with the increase of the number of pretext tasks. This paper introduces a method to select a group of pretext tasks among a set of candidates. The method we propose estimates calibrated weights for the partial losses corresponding to the considered pretext tasks during the self-supervised training process. The experiments conducted on automatic speech recognition, speaker and emotion recognition validate our approach, as the groups selected and weighted with our method perform better than classic baselines, thus facilitating the selection and combination of relevant pseudo-labels for self-supervised representation learning.

(0 min) After deploying a clinical prediction model, subsequently collected data can be used to fine-tune its predictions and adapt to temporal shifts. Because model updating carries risks of over-updating/fitting, we study online methods with performance guarantees. We introduce two procedures for continual recalibration or revision of an underlying prediction model: Bayesian logistic regression (BLR) and a Markov variant that explicitly models distribution shifts (MarBLR). We perform empirical evaluation via simulations and a real-world study predicting COPD risk. We derive "Type I and II" regret bounds, which guarantee the procedures are non-inferior to a static model and competitive with an oracle logistic reviser in terms of the average loss. Both procedures consistently outperformed the static model and other online logistic revision methods. In simulations, the average estimated calibration index (aECI) of the original model was 0.828 (95%CI 0.818-0.938). Online recalibration using BLR and MarBLR improved the aECI, attaining 0.265 (95%CI 0.230-0.300) and 0.241 (95%CI 0.216-0.266), respectively. When performing more extensive logistic model revisions, BLR and MarBLR increased the average AUC (aAUC) from 0.767 (95%CI 0.765-0.769) to 0.800 (95%CI 0.798-0.802) and 0.799 (95%CI 0.797-0.801), respectively, in stationary settings and protected against substantial model decay. In the COPD study, BLR and MarBLR dynamically combined the original model with a continually-refitted gradient boosted tree to achieve aAUCs of 0.924 (95%CI 0.913-0.935) and 0.925 (95%CI 0.914-0.935), compared to the static model's aAUC of 0.904 (95%CI 0.892-0.916). Despite its simplicity, BLR is highly competitive with MarBLR. MarBLR outperforms BLR when its prior better reflects the data. BLR and MarBLR can improve the transportability of clinical prediction models and maintain their performance over time.

(0 min) We propose a novel prediction interval (PI) method for uncertainty quantification, which addresses three major issues with the state-of-the-art PI methods. First, existing PI methods require retraining of neural networks (NNs) for every given confidence level and suffer from the crossing issue in calculating multiple PIs. Second, they usually rely on customized loss functions with extra sensitive hyperparameters for which fine tuning is required to achieve a well-calibrated PI. Third, they usually underestimate uncertainties of out-of-distribution (OOD) samples leading to over-confident PIs. Our PI3NN method calculates PIs from linear combinations of three NNs, each of which is independently trained using the standard mean squared error loss. The coefficients of the linear combinations are computed using root-finding algorithms to ensure tight PIs for a given confidence level. We theoretically prove that PI3NN can calculate PIs for a series of confidence levels without retraining NNs and it completely avoids the crossing issue. Additionally, PI3NN does not introduce any unusual hyperparameters resulting in a stable performance. Furthermore, we address OOD identification challenge by introducing an initialization scheme which provides reasonably larger PIs of the OOD samples than those of the in-distribution samples. Benchmark and real-world experiments show that our method outperforms several state-of-the-art approaches with respect to predictive uncertainty quality, robustness, and OOD samples identification.

(0 min) Deep learning has shown remarkable progress in a wide range of problems. However, efficient training of such models requires large-scale datasets, and getting annotations for such datasets can be challenging and costly. In this work, we explore the use of user-generated freely available labels from web videos for video understanding. We create a benchmark dataset consisting of around 2 million videos with associated user-generated annotations and other meta information. We utilize the collected dataset for action classification and demonstrate its usefulness with existing small-scale annotated datasets, UCF101 and HMDB51. We study different loss functions and two pretraining strategies, simple and self-supervised learning. We also show how a network pretrained on the proposed dataset can help against video corruption and label noise in downstream datasets. We present this as a benchmark dataset in noisy learning for video understanding. The dataset, code, and trained models will be publicly available for future research.

(0 min) We present a new financial framework where two families of RL-based agents representing the Liquidity Providers and Liquidity Takers learn simultaneously to satisfy their objective. Thanks to a parametrized reward formulation and the use of Deep RL, each group learns a shared policy able to generalize and interpolate over a wide range of behaviors. This is a step towards a fully RL-based market simulator replicating complex market conditions particularly suited to study the dynamics of the financial market under various scenarios.

(0 min) Estimating the motion of the camera together with the 3D structure of the scene from a monocular vision system is a complex task that often relies on the so-called scene rigidity assumption. When observing a dynamic environment, this assumption is violated which leads to an ambiguity between the ego-motion of the camera and the motion of the objects. To solve this problem, we present a self-supervised learning framework for 3D object motion field estimation from monocular videos. Our contributions are two-fold. First, we propose a two-stage projection pipeline to explicitly disentangle the camera ego-motion and the object motions with dynamics attention module, called DAM. Specifically, we design an integrated motion model that estimates the motion of the camera and object in the first and second warping stages, respectively, controlled by the attention module through a shared motion encoder. Second, we propose an object motion field estimation through contrastive sample consensus, called CSAC, taking advantage of weak semantic prior (bounding box from an object detector) and geometric constraints (each object respects the rigid body motion model). Experiments on KITTI, Cityscapes, and Waymo Open Dataset demonstrate the relevance of our approach and show that our method outperforms state-of-the-art algorithms for the tasks of self-supervised monocular depth estimation, object motion segmentation, monocular scene flow estimation, and visual odometry.

(0 min) Customer behavior is often assumed to follow weak rationality, which implies that adding a product to an assortment will not increase the choice probability of another product in that assortment. However, an increasing amount of research has revealed that customers are not necessarily rational when making decisions. In this paper, we propose a new nonparametric choice model that relaxes this assumption and can model a wider range of customer behavior, such as decoy effects between products. In this model, each customer type is associated with a binary decision tree, which represents a decision process for making a purchase based on checking for the existence of specific products in the assortment. Together with a probability distribution over customer types, we show that the resulting model -- a decision forest -- is able to represent any customer choice model, including models that are inconsistent with weak rationality. We theoretically characterize the depth of the forest needed to fit a data set of historical assortments and prove that with high probability, a forest whose depth scales logarithmically in the number of assortments is sufficient to fit most data sets. We also propose two practical algorithms -- one based on column generation and one based on random sampling -- for estimating such models from data. Using synthetic data and real transaction data exhibiting non-rational behavior, we show that the model outperforms both rational and non-rational benchmark models in out-of-sample predictive ability.

(0 min) In this paper, the authors investigate the Deep Sea Treasure (DST) problem as proposed by Vamplew et al. Through a number of proofs, the authors show the original DST problem to be quite basic, and not always representative of practical Multi-Objective Optimization problems. In an attempt to bring theory closer to practice, the authors propose an alternative, improved version of the DST problem, and prove that some of the properties that simplify the original DST problem no longer hold. The authors also provide a reference implementation and perform a comparison between their implementation, and other existing open-source implementations of the problem. Finally, the authors also provide a complete Pareto-front for their new DST problem.

(0 min) Contrastive learning (CL) is one of the most successful paradigms for self-supervised learning (SSL). In a principled way, it considers two augmented ``views'' of the same image as positive to be pulled closer, and all other images negative to be pushed further apart. However, behind the impressive success of CL-based techniques, their formulation often relies on heavy-computation settings, including large sample batches, extensive training epochs, etc. We are thus motivated to tackle these issues and aim at establishing a simple, efficient, and yet competitive baseline of contrastive learning. Specifically, we identify, from theoretical and empirical studies, a noticeable negative-positive-coupling (NPC) effect in the widely used cross-entropy (InfoNCE) loss, leading to unsuitable learning efficiency with respect to the batch size. Indeed the phenomenon tends to be neglected in that optimizing infoNCE loss with a small-size batch is effective in solving easier SSL tasks. By properly addressing the NPC effect, we reach a decoupled contrastive learning (DCL) objective function, significantly improving SSL efficiency. DCL can achieve competitive performance, requiring neither large batches in SimCLR, momentum encoding in MoCo, or large epochs. We demonstrate the usefulness of DCL in various benchmarks, while manifesting its robustness being much less sensitive to suboptimal hyperparameters. Notably, our approach achieves $66.9\%$ ImageNet top-1 accuracy using batch size 256 within 200 epochs pre-training, outperforming its baseline SimCLR by $5.1\%$. With further optimized hyperparameters, DCL can improve the accuracy to $68.2\%$. We believe DCL provides a valuable baseline for future contrastive learning-based SSL studies.

(0 min) Probabilistic estimation of cardiac electrophysiological model parameters serves an important step towards model personalization and uncertain quantification. The expensive computation associated with these model simulations, however, makes direct Markov Chain Monte Carlo (MCMC) sampling of the posterior probability density function (pdf) of model parameters computationally intensive. Approximated posterior pdfs resulting from replacing the simulation model with a computationally efficient surrogate, on the other hand, have seen limited accuracy. In this paper, we present a Bayesian active learning method to directly approximate the posterior pdf function of cardiac model parameters, in which we intelligently select training points to query the simulation model in order to learn the posterior pdf using a small number of samples. We integrate a generative model into Bayesian active learning to allow approximating posterior pdf of high-dimensional model parameters at the resolution of the cardiac mesh. We further introduce new acquisition functions to focus the selection of training points on better approximating the shape rather than the modes of the posterior pdf of interest. We evaluated the presented method in estimating tissue excitability in a 3D cardiac electrophysiological model in a range of synthetic and real-data experiments. We demonstrated its improved accuracy in approximating the posterior pdf compared to Bayesian active learning using regular acquisition functions, and substantially reduced computational cost in comparison to existing standard or accelerated MCMC sampling.

(0 min) We investigate the computational performance of Artificial Neural Networks (ANNs) in semi-nonparametric instrumental variables (NPIV) models of high dimensional covariates that are relevant to empirical work in economics. We focus on efficient estimation of and inference on expectation functionals (such as weighted average derivatives) and use optimal criterion-based procedures (sieve minimum distance or SMD) and novel efficient score-based procedures (ES). Both these procedures use ANN to approximate the unknown function. Then, we provide a detailed practitioner's recipe for implementing these two classes of estimators. This involves the choice of tuning parameters both for the unknown functions (that include conditional expectations) but also for the choice of estimation of the optimal weights in SMD and the Riesz representers used with the ES estimators. Finally, we conduct a large set of Monte Carlo experiments that compares the finite-sample performance in complicated designs that involve a large set of regressors (up to 13 continuous), and various underlying nonlinearities and covariate correlations. Some of the takeaways from our results include: 1) tuning and optimization are delicate especially as the problem is nonconvex; 2) various architectures of the ANNs do not seem to matter for the designs we consider and given proper tuning, ANN methods perform well; 3) stable inferences are more difficult to achieve with ANN estimators; 4) optimal SMD based estimators perform adequately; 5) there seems to be a gap between implementation and approximation theory. Finally, we apply ANN NPIV to estimate average price elasticity and average derivatives in two demand examples.

(0 min) Belonging to the family of Bayesian nonparametrics, Gaussian process (GP) based approaches have well-documented merits not only in learning over a rich class of nonlinear functions, but also in quantifying the associated uncertainty. However, most GP methods rely on a single preselected kernel function, which may fall short in characterizing data samples that arrive sequentially in time-critical applications. To enable {\it online} kernel adaptation, the present work advocates an incremental ensemble (IE-) GP framework, where an EGP meta-learner employs an {\it ensemble} of GP learners, each having a unique kernel belonging to a prescribed kernel dictionary. With each GP expert leveraging the random feature-based approximation to perform online prediction and model update with {\it scalability}, the EGP meta-learner capitalizes on data-adaptive weights to synthesize the per-expert predictions. Further, the novel IE-GP is generalized to accommodate time-varying functions by modeling structured dynamics at the EGP meta-learner and within each GP learner. To benchmark the performance of IE-GP and its dynamic variant in the adversarial setting where the modeling assumptions are violated, rigorous performance analysis has been conducted via the notion of regret, as the norm in online convex optimization. Last but not the least, online unsupervised learning for dimensionality reduction is explored under the novel IE-GP framework. Synthetic and real data tests demonstrate the effectiveness of the proposed schemes.

(0 min) Low earth orbit (LEO) satellite-assisted communications have been considered as one of key elements in beyond 5G systems to provide wide coverage and cost-efficient data services. Such dynamic space-terrestrial topologies impose exponential increase in the degrees of freedom in network management. In this paper, we address two practical issues for an over-loaded LEO-terrestrial system. The first challenge is how to efficiently schedule resources to serve the massive number of connected users, such that more data and users can be delivered/served. The second challenge is how to make the algorithmic solution more resilient in adapting to dynamic wireless environments.To address them, we first propose an iterative suboptimal algorithm to provide an offline benchmark. To adapt to unforeseen variations, we propose an enhanced meta-critic learning algorithm (EMCL), where a hybrid neural network for parameterization and the Wolpertinger policy for action mapping are designed in EMCL. The results demonstrate EMCL's effectiveness and fast-response capabilities in over-loaded systems and in adapting to dynamic environments compare to previous actor-critic and meta-learning methods.

(0 min) With little to no parallel data available for programming languages, unsupervised methods are well-suited to source code translation. However, the majority of unsupervised machine translation approaches rely on back-translation, a method developed in the context of natural language translation and one that inherently involves training on noisy inputs. Unfortunately, source code is highly sensitive to small changes; a single token can result in compilation failures or erroneous programs, unlike natural languages where small inaccuracies may not change the meaning of a sentence. To address this issue, we propose to leverage an automated unit-testing system to filter out invalid translations, thereby creating a fully tested parallel corpus. We found that fine-tuning an unsupervised model with this filtered data set significantly reduces the noise in the translations so-generated, comfortably outperforming the state-of-the-art for all language pairs studied. In particular, for Java $\to$ Python and Python $\to$ C++ we outperform the best previous methods by more than 16% and 24% respectively, reducing the error rate by more than 35%.

(0 min) Recent years have witnessed an explosion of science conspiracy videos on the Internet, challenging science epistemology and public understanding of science. Scholars have started to examine the persuasion techniques used in conspiracy messages such as uncertainty and fear yet, little is understood about the visual narratives, especially how visual narratives differ in videos that debunk conspiracies versus those that propagate conspiracies. This paper addresses this gap in understanding visual framing in conspiracy videos through analyzing millions of frames from conspiracy and counter-conspiracy YouTube videos using computational methods. We found that conspiracy videos tended to use lower color variance and brightness, especially in thumbnails and earlier parts of the videos. This paper also demonstrates how researchers can integrate textual and visual features in machine learning models to study conspiracies on social media and discusses the implications of computational modeling for scholars interested in studying visual manipulation in the digital era. The analysis of visual and textual features presented in this paper could be useful for future studies focused on designing systems to identify conspiracy content on the Internet.

(0 min) Seismic wave propagation forms the basis for most aspects of seismological research, yet solving the wave equation is a major computational burden that inhibits the progress of research. This is exacerbated by the fact that new simulations must be performed when the velocity structure or source location is perturbed. Here, we explore a prototype framework for learning general solutions using a recently developed machine learning paradigm called Neural Operator. A trained Neural Operator can compute a solution in negligible time for any velocity structure or source location. We develop a scheme to train Neural Operators on an ensemble of simulations performed with random velocity models and source locations. As Neural Operators are grid-free, it is possible to evaluate solutions on higher resolution velocity models than trained on, providing additional computational efficiency. We illustrate the method with the 2D acoustic wave equation and demonstrate the method's applicability to seismic tomography, using reverse mode automatic differentiation to compute gradients of the wavefield with respect to the velocity structure. The developed procedure is nearly an order of magnitude faster than using conventional numerical methods for full waveform inversion.

(0 min) With the increasing penetration of renewable energy, frequency response and its security are of significant concerns for reliable power system operations. Frequency-constrained unit commitment (FCUC) is proposed to address this challenge. Despite existing efforts in modeling frequency characteristics in unit commitment (UC), current strategies can only handle oversimplified low-order frequency response models and do not consider wide-range operating conditions. This paper presents a generic data-driven framework for FCUC under high renewable penetration. Deep neural networks (DNNs) are trained to predict the frequency response using real data or high-fidelity simulation data. Next, the DNN is reformulated as a set of mixed-integer linear constraints to be incorporated into the ordinary UC formulation. In the data generation phase, all possible power injections are considered, and a region-of-interests active sampling is proposed to include power injection samples with frequency nadirs closer to the UFLC threshold, which significantly enhances the accuracy of frequency constraints in FCUC. The proposed FCUC is verified on the the IEEE 39-bus system. Then, a full-order dynamic model simulation using PSS/E verifies the effectiveness of FCUC in frequency-secure generator commitments.

(0 min) Fine-tuning of large pre-trained image and language models on small customized datasets has become increasingly popular for improved prediction and efficient use of limited resources. Fine-tuning requires identification of best models to transfer-learn from and quantifying transferability prevents expensive re-training on all of the candidate models/tasks pairs. We show that the statistical problems with covariance estimation drive the poor performance of H-score [Bao et al., 2019] -- a common baseline for newer metrics -- and propose shrinkage-based estimator. This results in up to 80% absolute gain in H-score correlation performance, making it competitive with the state-of-the-art LogME measure by You et al. [2021]. Our shrinkage-based H-score is 3-55 times faster to compute compared to LogME. Additionally, we look into a less common setting of target (as opposed to source) task selection. We identify previously overlooked problems in such settings with different number of labels, class-imbalance ratios etc. for some recent metrics e.g., LEEP [Nguyen et al., 2020] that resulted in them being misrepresented as leading measures. We propose a correction and recommend measuring correlation performance against relative accuracy in such settings. We also outline the difficulties of comparing feature-dependent metrics, both supervised (e.g. H-score) and unsupervised measures (e.g., Maximum Mean Discrepancy [Long et al., 2015]), across source models/layers with different feature embedding dimension. We show that dimensionality reduction methods allow for meaningful comparison across models and improved performance of some of these measures. We investigate performance of 14 different supervised and unsupervised metrics and demonstrate that even unsupervised metrics can identify the leading models for domain adaptation. We support our findings with ~65,000 (fine-tuning trials) experiments.

(0 min) We examine counterfactual explanations for explaining the decisions made by model-based AI systems. The counterfactual approach we consider defines an explanation as a set of the system's data inputs that causally drives the decision (i.e., changing the inputs in the set changes the decision) and is irreducible (i.e., changing any subset of the inputs does not change the decision). We (1) demonstrate how this framework may be used to provide explanations for decisions made by general, data-driven AI systems that may incorporate features with arbitrary data types and multiple predictive models, and (2) propose a heuristic procedure to find the most useful explanations depending on the context. We then contrast counterfactual explanations with methods that explain model predictions by weighting features according to their importance (e.g., SHAP, LIME) and present two fundamental reasons why we should carefully consider whether importance-weight explanations are well-suited to explain system decisions. Specifically, we show that (i) features that have a large importance weight for a model prediction may not affect the corresponding decision, and (ii) importance weights are insufficient to communicate whether and how features influence decisions. We demonstrate this with several concise examples and three detailed case studies that compare the counterfactual approach with SHAP to illustrate various conditions under which counterfactual explanations explain data-driven decisions better than importance weights.

(0 min) Decision-making often requires accurate estimation of treatment effects from observational data. This is challenging as outcomes of alternative decisions are not observed and have to be estimated. Previous methods estimate outcomes based on unconfoundedness but neglect any constraints that unconfoundedness imposes on the outcomes. In this paper, we propose a novel regularization framework for estimating average treatment effects that exploits unconfoundedness. To this end, we formalize unconfoundedness as an orthogonality constraint, which ensures that the outcomes are orthogonal to the treatment assignment. This orthogonality constraint is then included in the loss function via a regularization. Based on our regularization framework, we develop deep orthogonal networks for unconfounded treatments (DONUT), which learn outcomes that are orthogonal to the treatment assignment. Using a variety of benchmark datasets for estimating average treatment effects, we demonstrate that DONUT outperforms the state-of-the-art substantially.

(0 min) We aim at advancing open-vocabulary object detection, which detects objects described by arbitrary text inputs. The fundamental challenge is the availability of training data. Existing object detection datasets only contain hundreds of categories, and it is costly to scale further. To overcome this challenge, we propose ViLD, a training method via Vision and Language knowledge Distillation. Our method distills the knowledge from a pretrained open-vocabulary image classification model (teacher) into a two-stage detector (student). Specifically, we use the teacher model to encode category texts and image regions of object proposals. Then we train a student detector, whose region embeddings of detected boxes are aligned with the text and image embeddings inferred by the teacher. We benchmark on LVIS by holding out all rare categories as novel categories not seen during training. ViLD obtains 16.1 mask AP$_r$, even outperforming the supervised counterpart by 3.8 with a ResNet-50 backbone. The model can directly transfer to other datasets without finetuning, achieving 72.2 AP$_{50}$, 36.6 AP and 11.8 AP on PASCAL VOC, COCO and Objects365, respectively. On COCO, ViLD outperforms previous SOTA by 4.8 on novel AP and 11.4 on overall AP.

(0 min) We study infinite horizon discounted Mean Field Control (MFC) problems with common noise through the lens of Mean Field Markov Decision Processes (MFMDP). We allow the agents to use actions that are randomized not only at the individual level but also at the level of the population. This common randomization allows us to establish connections between both closed-loop and open-loop policies for MFC and Markov policies for the MFMDP. In particular, we show that there exists an optimal closed-loop policy for the original MFC. Building on this framework and the notion of state-action value function, we then propose reinforcement learning (RL) methods for such problems, by adapting existing tabular and deep RL methods to the mean-field setting. The main difficulty is the treatment of the population state, which is an input of the policy and the value function. We provide convergence guarantees for tabular algorithms based on discretizations of the simplex. Neural network based algorithms are more suitable for continuous spaces and allow us to avoid discretizing the mean field state space. Numerical examples are provided.

(0 min) There is an increasing realization that algorithmic inductive biases are central in preventing overfitting; empirically, we often see a benign overfitting phenomenon in overparameterized settings for natural learning algorithms, such as stochastic gradient descent (SGD), where little to no explicit regularization has been employed. This work considers this issue in arguably the most basic setting: constant-stepsize SGD (with iterate averaging or tail averaging) for linear regression in the overparameterized regime. Our main result provides a sharp excess risk bound, stated in terms of the full eigenspectrum of the data covariance matrix, that reveals a bias-variance decomposition characterizing when generalization is possible: (i) the variance bound is characterized in terms of an effective dimension (specific for SGD) and (ii) the bias bound provides a sharp geometric characterization in terms of the location of the initial iterate (and how it aligns with the data covariance matrix). More specifically, for SGD with iterate averaging, we demonstrate the sharpness of the established excess risk bound by proving a matching lower bound (up to constant factors). For SGD with tail averaging, we show its advantage over SGD with iterate averaging by proving a better excess risk bound together with a nearly matching lower bound. Moreover, we reflect on a number of notable differences between the algorithmic regularization afforded by (unregularized) SGD in comparison to ordinary least squares (minimum-norm interpolation) and ridge regression. Experimental results on synthetic data corroborate our theoretical findings.

(0 min) Despite their recent popularity and well known advantages, dense retrievers still lag behind sparse methods such as BM25 in their ability to reliably match salient phrases and rare entities in the query. It has been argued that this is an inherent limitation of dense models. We disprove this claim by introducing the Salient Phrase Aware Retriever (SPAR), a dense retriever with the lexical matching capacity of a sparse model. In particular, we show that a dense retriever {\Lambda} can be trained to imitate a sparse one, and SPAR is built by augmenting a standard dense retriever with {\Lambda}. When evaluated on five open-domain question answering datasets and the MS MARCO passage retrieval task, SPAR sets a new state of the art for dense and sparse retrievers and can match or exceed the performance of more complicated dense-sparse hybrid systems.

(0 min) 3D object detection often involves complicated training and testing pipelines, which require substantial domain knowledge about individual datasets. Inspired by recent non-maximum suppression-free 2D object detection models, we propose a 3D object detection architecture on point clouds. Our method models 3D object detection as message passing on a dynamic graph, generalizing the DGCNN framework to predict a set of objects. In our construction, we remove the necessity of post-processing via object confidence aggregation or non-maximum suppression. To facilitate object detection from sparse point clouds, we also propose a set-to-set distillation approach customized to 3D detection. This approach aligns the outputs of the teacher model and the student model in a permutation-invariant fashion, significantly simplifying knowledge distillation for the 3D detection task. Our method achieves state-of-the-art performance on autonomous driving benchmarks. We also provide abundant analysis of the detection model and distillation framework.

(0 min) Graph Convolutional Representation (GCR) has achieved impressive performance for graph data representation. However, existing GCR is generally defined on the input fixed graph which may restrict the representation capacity and also be vulnerable to the structural attacks and noises. To address this issue, we propose a novel Latent Graph Convolutional Representation (LatGCR) for robust graph data representation and learning. Our LatGCR is derived based on reformulating graph convolutional representation from the aspect of graph neighborhood reconstruction. Given an input graph $\textbf{A}$, LatGCR aims to generate a flexible latent graph $\widetilde{\textbf{A}}$ for graph convolutional representation which obviously enhances the representation capacity and also performs robustly w.r.t graph structural attacks and noises. Moreover, LatGCR is implemented in a self-supervised manner and thus provides a basic block for both supervised and unsupervised graph learning tasks. Experiments on several datasets demonstrate the effectiveness and robustness of LatGCR.

(0 min) In this paper, we study the statistical limits of deep learning techniques for solving elliptic partial differential equations (PDEs) from random samples using the Deep Ritz Method (DRM) and Physics-Informed Neural Networks (PINNs). To simplify the problem, we focus on a prototype elliptic PDE: the Schr\"odinger equation on a hypercube with zero Dirichlet boundary condition, which has wide application in the quantum-mechanical systems. We establish upper and lower bounds for both methods, which improves upon concurrently developed upper bounds for this problem via a fast rate generalization bound. We discover that the current Deep Ritz Methods is sub-optimal and propose a modified version of it. We also prove that PINN and the modified version of DRM can achieve minimax optimal bounds over Sobolev spaces. Empirically, following recent work which has shown that the deep model accuracy will improve with growing training sets according to a power law, we supply computational experiments to show a similar behavior of dimension dependent power law for deep PDE solvers.

(0 min) Communications standards are designed via committees of humans holding repeated meetings over months or even years until consensus is achieved. This includes decisions regarding the modulation and coding schemes to be supported over an air interface. We propose a way to "automate" the selection of the set of modulation and coding schemes to be supported over a given air interface and thereby streamline both the standards design process and the ease of extending the standard to support new modulation schemes applicable to new higher-level applications and services. Our scheme involves machine learning, whereby a constructor entity submits proposals to an evaluator entity, which returns a score for the proposal. The constructor employs reinforcement learning to iterate on its submitted proposals until a score is achieved that was previously agreed upon by both constructor and evaluator to be indicative of satisfying the required design criteria (including performance metrics for transmissions over the interface).

(0 min) We introduce a framework for multi-camera 3D object detection. In contrast to existing works, which estimate 3D bounding boxes directly from monocular images or use depth prediction networks to generate input for 3D object detection from 2D information, our method manipulates predictions directly in 3D space. Our architecture extracts 2D features from multiple camera images and then uses a sparse set of 3D object queries to index into these 2D features, linking 3D positions to multi-view images using camera transformation matrices. Finally, our model makes a bounding box prediction per object query, using a set-to-set loss to measure the discrepancy between the ground-truth and the prediction. This top-down approach outperforms its bottom-up counterpart in which object bounding box prediction follows per-pixel depth estimation, since it does not suffer from the compounding error introduced by a depth prediction model. Moreover, our method does not require post-processing such as non-maximum suppression, dramatically improving inference speed. We achieve state-of-the-art performance on the nuScenes autonomous driving benchmark.

(0 min) Blowfish privacy is a recent generalisation of differential privacy that enables improved utility while maintaining privacy policies with semantic guarantees, a factor that has driven the popularity of differential privacy in computer science. This paper relates Blowfish privacy to an important measure of privacy loss of information channels from the communications theory community: min-entropy leakage. Symmetry in an input data neighbouring relation is central to known connections between differential privacy and min-entropy leakage. But while differential privacy exhibits strong symmetry, Blowfish neighbouring relations correspond to arbitrary simple graphs owing to the framework's flexible privacy policies. To bound the min-entropy leakage of Blowfish-private mechanisms we organise our analysis over symmetrical partitions corresponding to orbits of graph automorphism groups. A construction meeting our bound with asymptotic equality demonstrates tightness.

(0 min) Our hands reveal important information such as the pulsing of our veins which help us determine the blood pressure, tremors indicative of motor control, or neurodegenerative disorders such as Essential Tremor or Parkinson's disease. The Computerized Classification of Micro-Motions in the hand using waveforms from mobile phone videos is a novel method that uses Eulerian Video Magnification, Skeletonization, Heatmapping, and the kNN machine learning model to detect the micro-motions in the human hand, synthesize their waveforms, and classify these. The pre-processing is achieved by using Eulerian Video Magnification, Skeletonization, and Heat-mapping to magnify the micro-motions, landmark essential features of the hand, and determine the extent of motion, respectively. Following pre-processing, the visible motions are manually labeled by appropriately grouping pixels to represent a particular label correctly. These labeled motions of the pixels are converted into waveforms. Finally, these waveforms are classified into four categories - hand or finger movements, vein movement, background motion, and movement of the rest of the body due to respiration using the kNN model. The final accuracy obtained was around 92 percent.

(0 min) Humans have mental models that allow them to plan, experiment, and reason in the physical world. How should an intelligent agent go about learning such models? In this paper, we will study if models of the world learned in an open-ended physics environment, without any specific tasks, can be reused for downstream physics reasoning tasks. To this end, we build a benchmark Open-ended Physics ENvironment (OPEn) and also design several tasks to test learning representations in this environment explicitly. This setting reflects the conditions in which real agents (i.e. rolling robots) find themselves, where they may be placed in a new kind of environment and must adapt without any teacher to tell them how this environment works. This setting is challenging because it requires solving an exploration problem in addition to a model building and representation learning problem. We test several existing RL-based exploration methods on this benchmark and find that an agent using unsupervised contrastive learning for representation learning, and impact-driven learning for exploration, achieved the best results. However, all models still fall short in sample efficiency when transferring to the downstream tasks. We expect that OPEn will encourage the development of novel rolling robot agents that can build reusable mental models of the world that facilitate many tasks.

(0 min) Using observational data to estimate the effect of a treatment is a powerful tool for decision-making when randomized experiments are infeasible or costly. However, observational data often yields biased estimates of treatment effects, since treatment assignment can be confounded by unobserved variables. A remedy is offered by deconfounding methods that adjust for such unobserved confounders. In this paper, we develop the Sequential Deconfounder, a method that enables estimating individualized treatment effects over time in presence of unobserved confounders. This is the first deconfounding method that can be used in a general sequential setting (i.e., with one or more treatments assigned at each timestep). The Sequential Deconfounder uses a novel Gaussian process latent variable model to infer substitutes for the unobserved confounders, which are then used in conjunction with an outcome model to estimate treatment effects over time. We prove that using our method yields unbiased estimates of individualized treatment responses over time. Using simulated and real medical data, we demonstrate the efficacy of our method in deconfounding the estimation of treatment responses over time.

(0 min) Survival analysis is a critical tool for the modelling of time-to-event data, such as life expectancy after a cancer diagnosis or optimal maintenance scheduling for complex machinery. However, current neural network models provide an imperfect solution for survival analysis as they either restrict the shape of the target probability distribution or restrict the estimation to pre-determined times. As a consequence, current survival neural networks lack the ability to estimate a generic function without prior knowledge of its structure. In this article, we present the metaparametric neural network framework that encompasses existing survival analysis methods and enables their extension to solve the aforementioned issues. This framework allows survival neural networks to satisfy the same independence of generic function estimation from the underlying data structure that characterizes their regression and classification counterparts. Further, we demonstrate the application of the metaparametric framework using both simulated and large real-world datasets and show that it outperforms the current state-of-the-art methods in (i) capturing nonlinearities, and (ii) identifying temporal patterns, leading to more accurate overall estimations whilst placing no restrictions on the underlying function structure.

(0 min) Online auction has been very widespread in the recent years. Platform administrators are working hard to refine their auction mechanisms that will generate high profits while maintaining a fair resource allocation. With the advancement of computing technology and the bottleneck in theoretical frameworks, researchers are shifting gears towards online auction designs using deep learning approaches. In this article, we summarized some common deep learning infrastructures adopted in auction mechanism designs and showed how these architectures are evolving. We also discussed how researchers are tackling with the constraints and concerns in the large and dynamic industrial settings. Finally, we pointed out several currently unresolved issues for future directions.

(0 min) Neural Radiance Fields (NeRF) have recently demonstrated photo-realistic results for the task of novel view synthesis. In this paper, we propose to apply novel view synthesis to the robot relocalization problem: we demonstrate improvement of camera pose regression thanks to an additional synthetic dataset rendered by the NeRF class of algorithm. To avoid spawning novel views in irrelevant places we selected virtual camera locations from NeRF internal representation of the 3D geometry of the scene. We further improved localization accuracy of pose regressors using synthesized realistic and geometry consistent images as data augmentation during training. At the time of publication, our approach improved state of the art with a 60% lower error on Cambridge Landmarks and 7-scenes datasets. Hence, the resulting accuracy becomes comparable to structure-based methods, without any architecture modification or domain adaptation constraints. Since our method allows almost infinite generation of training data, we investigated limitations of camera pose regression depending on size and distribution of data used for training on public benchmarks. We concluded that pose regression accuracy is mostly bounded by relatively small and biased datasets rather than capacity of the pose regression model to solve the localization task.

(0 min) Learning generative object models from unlabelled videos is a long standing problem and required for causal scene modeling. We decompose this problem into three easier subtasks, and provide candidate solutions for each of them. Inspired by the Common Fate Principle of Gestalt Psychology, we first extract (noisy) masks of moving objects via unsupervised motion segmentation. Second, generative models are trained on the masks of the background and the moving objects, respectively. Third, background and foreground models are combined in a conditional "dead leaves" scene model to sample novel scene configurations where occlusions and depth layering arise naturally. To evaluate the individual stages, we introduce the Fishbowl dataset positioned between complex real-world scenes and common object-centric benchmarks of simplistic objects. We show that our approach allows learning generative models that generalize beyond the occlusions present in the input videos, and represent scenes in a modular fashion that allows sampling plausible scenes outside the training distribution by permitting, for instance, object numbers or densities not observed in the training set.

(0 min) Amongst others, the adoption of Rectified Linear Units (ReLUs) is regarded as one of the ingredients of the success of deep learning. ReLU activation has been shown to mitigate the vanishing gradient issue, to encourage sparsity in the learned parameters, and to allow for efficient backpropagation. In this paper, we recognize that the non-linear behavior of the ReLU function gives rise to a natural clustering when the pattern of active neurons is considered. This observation helps to deepen the learning mechanism of the network; in fact, we demonstrate that, within each cluster, the network can be fully represented as an affine map. The consequence is that we are able to recover an explanation, in the form of feature importance, for the predictions done by the network to the instances belonging to the cluster. Therefore, the methodology we propose is able to increase the level of interpretability of a fully connected feedforward ReLU neural network, downstream from the fitting phase of the model, without altering the structure of the network. A simulation study and the empirical application to the Titanic dataset, show the capability of the method to bridge the gap between the algorithm optimization and the human understandability of the black box deep ReLU networks.

(0 min) We investigate the ability to detect slag formations in images from inside a Grate-Kiln system furnace with two deep convolutional neural networks. The conditions inside the furnace cause occasional obstructions of the camera view. Our approach suggests dealing with this problem by introducing a convLSTM-layer in the deep convolutional neural network. The results show that it is possible to achieve sufficient performance to automate the decision of timely countermeasures in the industrial operational setting. Furthermore, the addition of the convLSTM-layer results in fewer outlying predictions and a lower running variance of the fraction of detected slag in the image time series.

(0 min) We present a proof-of-concept technique for the inverse design of electromagnetic devices motivated by the policy gradient method in reinforcement learning, named PHORCED (PHotonic Optimization using REINFORCE Criteria for Enhanced Design). This technique uses a probabilistic generative neural network interfaced with an electromagnetic solver to assist in the design of photonic devices, such as grating couplers. We show that PHORCED obtains better performing grating coupler designs than local gradient-based inverse design via the adjoint method, while potentially providing faster convergence over competing state-of-the-art generative methods. As a further example of the benefits of this method, we implement transfer learning with PHORCED, demonstrating that a neural network trained to optimize 8$^\circ$ grating couplers can then be re-trained on grating couplers with alternate scattering angles while requiring >10$\times$ fewer simulations than control cases.

(0 min) Local feature attribution methods are increasingly used to explain complex machine learning models. However, current methods are limited because they are extremely expensive to compute or are not capable of explaining a distributed series of models where each model is owned by a separate institution. The latter is particularly important because it often arises in finance where explanations are mandated. Here, we present DeepSHAP, a tractable method to propagate local feature attributions through complex series of models based on a connection to the Shapley value. We evaluate DeepSHAP across biological, health, and financial datasets to show that it provides equally salient explanations an order of magnitude faster than existing model-agnostic attribution techniques and demonstrate its use in an important distributed series of models setting.

(0 min) In this contribution, we investigate the effectiveness of deep fusion of text and audio features for categorical and dimensional speech emotion recognition (SER). We propose a novel, multistage fusion method where the two information streams are integrated in several layers of a deep neural network (DNN), and contrast it with a single-stage one where the streams are merged in a single point. Both methods depend on extracting summary linguistic embeddings from a pre-trained BERT model, and conditioning one or more intermediate representations of a convolutional model operating on log-Mel spectrograms. Experiments on the widely used IEMOCAP and MSP-Podcast databases demonstrate that the two fusion methods clearly outperform a shallow (late) fusion baseline and their unimodal constituents, both in terms of quantitative performance and qualitative behaviour. Our accompanying analysis further reveals a hitherto unexplored role of the underlying dialogue acts on unimodal and bimodal SER, with different models showing a biased behaviour across different acts. Overall, our multistage fusion shows better quantitative performance, surpassing all alternatives on most of our evaluations. This illustrates the potential of multistage fusion in better assimilating text and audio information.

(0 min) Recent studies in singing voice synthesis have achieved high-quality results leveraging advances in text-to-speech models based on deep neural networks. One of the main issues in training singing voice synthesis models is that they require melody and lyric labels to be temporally aligned with audio data. The temporal alignment is a time-exhausting manual work in preparing for the training data. To address the issue, we propose a melody-unsupervision model that requires only audio-and-lyrics pairs without temporal alignment in training time but generates singing voice audio given a melody and lyrics input in inference time. The proposed model is composed of a phoneme classifier and a singing voice generator jointly trained in an end-to-end manner. The model can be fine-tuned by adjusting the amount of supervision with temporally aligned melody labels. Through experiments in melody-unsupervision and semi-supervision settings, we compare the audio quality of synthesized singing voice. We also show that the proposed model is capable of being trained with speech audio and text labels but can generate singing voice in inference time.

(0 min) We present a data-driven approach to characterizing nonidentifiability of a model's parameters and illustrate it through dynamic kinetic models. By employing Diffusion Maps and their extensions, we discover the minimal combinations of parameters required to characterize the dynamic output behavior: a set of effective parameters for the model. Furthermore, we use Conformal Autoencoder Neural Networks, as well as a kernel-based Jointly Smooth Function technique, to disentangle the redundant parameter combinations that do not affect the output behavior from the ones that do. We discuss the interpretability of our data-driven effective parameters and demonstrate the utility of the approach both for behavior prediction and parameter estimation. In the latter task, it becomes important to describe level sets in parameter space that are consistent with a particular output behavior. We validate our approach on a model of multisite phosphorylation, where a reduced set of effective parameters, nonlinear combinations of the physical ones, has previously been established analytically.

(0 min) Classification of images within the compressed domain offers significant benefits. These benefits include reduced memory and computational requirements of a classification system. This paper proposes two such methods as a proof of concept: The first classifies within the JPEG image transform domain (i.e. DCT transform data); the second classifies the JPEG compressed binary bitstream directly. These two methods are implemented using Residual Network CNNs and an adapted Vision Transformer. Top-1 accuracy of approximately 70% and 60% were achieved using these methods respectively when classifying the Caltech C101 database. Although these results are significantly behind the state of the art for classification for this database (~95%), it illustrates the first time direct bitstream image classification has been achieved. This work confirms that direct bitstream image classification is possible and could be utilised in a first pass database screening of a raw bitstream (within a wired or wireless network) or where computational, memory and bandwidth requirements are severely restricted.

(0 min) Background: Machine learning-based security detection models have become prevalent in modern malware and intrusion detection systems. However, previous studies show that such models are susceptible to adversarial evasion attacks. In this type of attack, inputs (i.e., adversarial examples) are specially crafted by intelligent malicious adversaries, with the aim of being misclassified by existing state-of-the-art models (e.g., deep neural networks). Once the attackers can fool a classifier to think that a malicious input is actually benign, they can render a machine learning-based malware or intrusion detection system ineffective. Goal: To help security practitioners and researchers build a more robust model against non-adaptive, white-box, and non-targeted adversarial evasion attacks through the idea of an ensemble model. Method: We propose an approach called Omni, the main idea of which is to explore methods that create an ensemble of "unexpected models"; i.e., models whose control hyperparameters have a large distance to the hyperparameters of an adversary's target model, with which we then make an optimized weighted ensemble prediction. Result: In studies with five types of adversarial evasion attacks (FGSM, BIM, JSMA, DeepFooland Carlini-Wagner) on five security datasets (NSL-KDD, CIC-IDS-2017, CSE-CIC-IDS2018, CICAnd-Mal2017, and the Contagio PDF dataset), we show Omni is a promising approach as a defense strategy against adversarial attacks when compared with other baseline treatments. Conclusion: When employing ensemble defense against adversarial evasion attacks, we suggest creating an ensemble with unexpected models that are distant from the attacker's expected model (i.e., target model) through methods such as hyperparameter optimization.

(0 min) We present extensive empirical evidence showing that current Bayesian simulation-based inference algorithms are inadequate for the falsificationist methodology of scientific inquiry. Our results collected through months of experimental computations show that all benchmarked algorithms -- (S)NPE, (S)NRE, SNL and variants of ABC -- may produce overconfident posterior approximations, which makes them demonstrably unreliable and dangerous if one's scientific goal is to constrain parameters of interest. We believe that failing to address this issue will lead to a well-founded trust crisis in simulation-based inference. For this reason, we argue that research efforts should now consider theoretical and methodological developments of conservative approximate inference algorithms and present research directions towards this objective. In this regard, we show empirical evidence that ensembles are consistently more reliable.

(0 min) Dictionary learning aims to find a dictionary under which the training data can be sparsely represented, and it is usually achieved by iteratively applying two stages: sparse coding and dictionary update. Typical methods for dictionary update focuses on refining both dictionary atoms and their corresponding sparse coefficients by using the sparsity patterns obtained from sparse coding stage, and hence it is a non-convex bilinear inverse problem. In this paper, we propose a Rank-One Matrix Decomposition (ROMD) algorithm to recast this challenge into a convex problem by resolving these two variables into a set of rank-one matrices. Different from methods in the literature, ROMD updates the whole dictionary at a time using convex programming. The advantages hence include both convergence guarantees for dictionary update and faster convergence of the whole dictionary learning. The performance of ROMD is compared with other benchmark dictionary learning algorithms. The results show the improvement of ROMD in recovery accuracy, especially in the cases of high sparsity level and fewer observation data.

(0 min) Single channel blind source separation (SCBSS) refers to separate multiple sources from a mixed signal collected by a single sensor. The existing methods for SCBSS mainly focus on separating two sources and have weak generalization performance. To address these problems, an algorithm is proposed in this paper to separate multiple sources from a mixture by designing a parallel dual generative adversarial Network (PDualGAN) that can build the relationship between a mixture and the corresponding multiple sources to realize one-to-multiple cross-domain mapping. This algorithm can be applied to any mixed model such as linear instantaneous mixed model and convolutional mixed model. Besides, one-to-multiple datasets are created which including the mixtures and corresponding sources for this study. The experiment was carried out on four different datasets and tested with signals mixed in different proportions. Experimental results show that the proposed algorithm can achieve high performance in peak signal-to-noise ratio (PSNR) and correlation, which outperforms state-of-the-art algorithms.

(0 min) The adversarial attack literature contains a myriad of algorithms for crafting perturbations which yield pathological behavior in neural networks. In many cases, multiple algorithms target the same tasks and even enforce the same constraints. In this work, we show that different attack algorithms produce adversarial examples which are distinct not only in their effectiveness but also in how they qualitatively affect their victims. We begin by demonstrating that one can determine the attack algorithm that crafted an adversarial example. Then, we leverage recent advances in parameter-space saliency maps to show, both visually and quantitatively, that adversarial attack algorithms differ in which parts of the network and image they target. Our findings suggest that prospective adversarial attacks should be compared not only via their success rates at fooling models but also via deeper downstream effects they have on victims.

(0 min) During the COVID-19 pandemic, most of the human-to-human interactions have been stopped. To mitigate the spread of deadly coronavirus, many offices took the initiative so that the employees can work from home. But, tracking the employees and finding out if they are really performing what they were supposed to turn out to be a serious challenge for all the companies and organizations who are facilitating "Work From Home". To deal with the challenge effectively, we came up with a solution to track the employees with face recognition. We have been testing this system experimentally for our office. To train the face recognition module, we used FaceNet with KNN using the Labeled Faces in the Wild (LFW) dataset and achieved 97.8\% accuracy. We integrated the trained model into our central system, where the employees log their time. In this paper, we discuss in brief the system we have been experimenting with and the pros and cons of the system.

(0 min) Linear temporal logic (LTL) is a specification language for finite sequences (called traces) widely used in program verification, motion planning in robotics, process mining, and many other areas. We consider the problem of learning LTL formulas for classifying traces; despite a growing interest of the research community, existing solutions suffer from two limitations: they do not scale beyond small formulas, and they may exhaust computational resources without returning any result. We introduce a new algorithm addressing both issues: our algorithm is able to construct formulas an order of magnitude larger than previous methods, and it is anytime, meaning that it in most cases successfully outputs a formula, albeit possibly not of minimal size. We evaluate the performances of our algorithm using an open source implementation against publicly available benchmarks.

(0 min) The abundance of data has given machine learning considerable momentum in natural sciences and engineering. However, the modeling of simulated physical processes remains difficult. A key problem is the correct handling of geometric boundaries. While triangularized geometric boundaries are very common in engineering applications, they are notoriously difficult to model by machine learning approaches due to their heterogeneity with respect to size and orientation. In this work, we introduce Boundary Graph Neural Networks (BGNNs), which dynamically modify graph structures to address boundary conditions. Boundary graph structures are constructed via modifying edges, augmenting node features, and dynamically inserting virtual nodes. The new BGNNs are tested on complex 3D granular flow processes of hoppers and rotating drums which are standard components of industrial machinery. Using precise simulations that are obtained by an expensive and complex discrete element method, BGNNs are evaluated in terms of computational efficiency as well as prediction accuracy of particle flows and mixing entropies. Even if complex boundaries are present, BGNNs are able to accurately reproduce 3D granular flows within simulation uncertainties over hundreds of thousands of simulation timesteps, and most notably particles completely stay within the geometric objects without using handcrafted conditions or restrictions.

(0 min) Machine learning image classifiers are susceptible to adversarial and corruption perturbations. Adding imperceptible noise to images can lead to severe misclassifications of the machine learning model. Using $L_p$-norms for measuring the size of the noise fails to capture human similarity perception, which is why optimal transport based distance measures like the Wasserstein metric are increasingly being used in the field of adversarial robustness. Verifying the robustness of classifiers using the Wasserstein metric can be achieved by proving the absence of adversarial examples (certification) or proving their presence (attack). In this work we present a framework based on the work by Levine and Feizi, which allows us to transfer existing certification methods for convex polytopes or $L_1$-balls to the Wasserstein threat model. The resulting certification can be complete or incomplete, depending on whether convex polytopes or $L_1$-balls were chosen. Additionally, we present a new Wasserstein adversarial attack that is projected gradient descent based and which has a significantly reduced computational burden compared to existing attack approaches.

(0 min) Existing generative pre-trained language models (e.g., GPT) focus on modeling the language structure and semantics of general texts. However, those models do not consider the numerical properties of numbers and cannot perform robustly on numerical reasoning tasks (e.g., math word problems and measurement estimation). In this paper, we propose NumGPT, a generative pre-trained model that explicitly models the numerical properties of numbers in texts. Specifically, it leverages a prototype-based numeral embedding to encode the mantissa of the number and an individual embedding to encode the exponent of the number. A numeral-aware loss function is designed to integrate numerals into the pre-training objective of NumGPT. We conduct extensive experiments on four different datasets to evaluate the numeracy ability of NumGPT. The experiment results show that NumGPT outperforms baseline models (e.g., GPT and GPT with DICE) on a range of numerical reasoning tasks such as measurement estimation, number comparison, math word problems, and magnitude classification. Ablation studies are also conducted to evaluate the impact of pre-training and model hyperparameters on the performance.

(0 min) The latest Deep Learning (DL) models for detection and classification have achieved an unprecedented performance over classical machine learning algorithms. However, DL models are black-box methods hard to debug, interpret, and certify. DL alone cannot provide explanations that can be validated by a non technical audience. In contrast, symbolic AI systems that convert concepts into rules or symbols -- such as knowledge graphs -- are easier to explain. However, they present lower generalisation and scaling capabilities. A very important challenge is to fuse DL representations with expert knowledge. One way to address this challenge, as well as the performance-explainability trade-off is by leveraging the best of both streams without obviating domain expert knowledge. We tackle such problem by considering the symbolic knowledge is expressed in form of a domain expert knowledge graph. We present the eXplainable Neural-symbolic learning (X-NeSyL) methodology, designed to learn both symbolic and deep representations, together with an explainability metric to assess the level of alignment of machine and human expert explanations. The ultimate objective is to fuse DL representations with expert domain knowledge during the learning process to serve as a sound basis for explainability. X-NeSyL methodology involves the concrete use of two notions of explanation at inference and training time respectively: 1) EXPLANet: Expert-aligned eXplainable Part-based cLAssifier NETwork Architecture, a compositional CNN that makes use of symbolic representations, and 2) SHAP-Backprop, an explainable AI-informed training procedure that guides the DL process to align with such symbolic representations in form of knowledge graphs. We showcase X-NeSyL methodology using MonuMAI dataset for monument facade image classification, and demonstrate that our approach improves explainability and performance.

(0 min) Embodied conversational agents benefit from being able to accompany their speech with gestures. Although many data-driven approaches to gesture generation have been proposed in recent years, it is still unclear whether such systems can consistently generate gestures that convey meaning. We investigate which gesture properties (phase, category, and semantics) can be predicted from speech text and/or audio using contemporary deep learning. In extensive experiments, we show that gesture properties related to gesture meaning (semantics and category) are predictable from text features (time-aligned FastText embeddings) alone, but not from prosodic audio features, while rhythm-related gesture properties (phase) on the other hand can be predicted from audio features better than from text. These results are encouraging as they indicate that it is possible to equip an embodied agent with content-wise meaningful co-speech gestures using a machine-learning model.

(0 min) Robots in our daily surroundings are increasing day by day. Their usability and acceptability largely depend on their explicit and implicit interaction capability with fellow human beings. As a result, social behavior is one of the most sought-after qualities that a robot can possess. However, there is no specific aspect and/or feature that defines socially acceptable behavior and it largely depends on the situation, application, and society. In this article, we investigate one such social behavior for collocated robots. Imagine a group of people is interacting with each other and we want to join the group. We as human beings do it in a socially acceptable manner, i.e., within the group, we do position ourselves in such a way that we can participate in the group activity without disturbing/obstructing anybody. To possess such a quality, first, a robot needs to determine the formation of the group and then determine a position for itself, which we humans do implicitly. The theory of f-formation can be utilized for this purpose. As the types of formations can be very diverse, detecting the social groups is not a trivial task. In this article, we provide a comprehensive survey of the existing work on social interaction and group detection using f-formation for robotics and other applications. We also put forward a novel holistic survey framework combining all the possible concerns and modules relevant to this problem. We define taxonomies based on methods, camera views, datasets, detection capabilities and scale, evaluation approaches, and application areas. We discuss certain open challenges and limitations in current literature along with possible future research directions based on this framework. In particular, we discuss the existing methods/techniques and their relative merits and demerits, applications, and provide a set of unsolved but relevant problems in this domain.

(0 min) Recently, Zhang et al. (2021) developed a new neural network architecture based on $\ell_\infty$-distance functions, which naturally possesses certified robustness by its construction. Despite the excellent theoretical properties, the model so far can only achieve comparable performance to conventional networks. In this paper, we significantly boost the certified robustness of $\ell_\infty$-distance nets through a careful analysis of its training process. In particular, we show the $\ell_p$-relaxation, a crucial way to overcome the non-smoothness of the model, leads to an unexpected large Lipschitz constant at the early training stage. This makes the optimization insufficient using hinge loss and produces sub-optimal solutions. Given these findings, we propose a simple approach to address the issues above by using a novel objective function that combines a scaled cross-entropy loss with clipped hinge loss. Our experiments show that using the proposed training strategy, the certified accuracy of $\ell_\infty$-distance net can be dramatically improved from 33.30% to 40.06% on CIFAR-10 ($\epsilon=8/255$), meanwhile significantly outperforming other approaches in this area. Such a result clearly demonstrates the effectiveness and potential of $\ell_\infty$-distance net for certified robustness.

(0 min) Discovering successful coordinated behaviors is a central challenge in Multi-Agent Reinforcement Learning (MARL) since it requires exploring a joint action space that grows exponentially with the number of agents. In this paper, we propose a mechanism for achieving sufficient exploration and coordination in a team of agents. Specifically, agents are rewarded for contributing to a more diversified team behavior by employing proper intrinsic motivation functions. To learn meaningful coordination protocols, we structure agents' interactions by introducing a novel framework, where at each timestep, an agent simulates counterfactual rollouts of its policy and, through a sequence of computations, assesses the gap between other agents' current behaviors and their targets. Actions that minimize the gap are considered highly influential and are rewarded. We evaluate our approach on a set of challenging tasks with sparse rewards and partial observability that require learning complex cooperative strategies under a proper exploration scheme, such as the StarCraft Multi-Agent Challenge. Our methods show significantly improved performances over different baselines across all tasks.

(0 min) The COVID-19 pandemic raises the problem of adapting face recognition systems to the new reality, where people may wear surgical masks to cover their noses and mouths. Traditional data sets (e.g., CelebA, CASIA-WebFace) used for training these systems were released before the pandemic, so they now seem unsuited due to the lack of examples of people wearing masks. We propose a method for enhancing data sets containing faces without masks by creating synthetic masks and overlaying them on faces in the original images. Our method relies on SparkAR Studio, a developer program made by Facebook that is used to create Instagram face filters. In our approach, we use 9 masks of different colors, shapes and fabrics. We employ our method to generate a number of 445,446 (90%) samples of masks for the CASIA-WebFace data set and 196,254 (96.8%) masks for the CelebA data set, releasing the mask images at https://github.com/securifai/masked_faces. We show that our method produces significantly more realistic training examples of masks overlaid on faces by asking volunteers to qualitatively compare it to other methods or data sets designed for the same task. We also demonstrate the usefulness of our method by evaluating state-of-the-art face recognition systems (FaceNet, VGG-face, ArcFace) trained on our enhanced data sets and showing that they outperform equivalent systems trained on original data sets (containing faces without masks) or competing data sets (containing masks generated by related methods), when the test benchmarks contain masked faces.

(0 min) Training deep neural networks may be challenging in real world data. Using models as black-boxes, even with transfer learning, can result in poor generalization or inconclusive results when it comes to small datasets or specific applications. This tutorial covers the basic steps as well as more recent options to improve models, in particular, but not restricted to, supervised learning. It can be particularly useful in datasets that are not as well-prepared as those in challenges, and also under scarce annotation and/or small data. We describe basic procedures: as data preparation, optimization and transfer learning, but also recent architectural choices such as use of transformer modules, alternative convolutional layers, activation functions, wide and deep networks, as well as training procedures including as curriculum, contrastive and self-supervised learning.

(0 min) While significant research advances have been made in the field of deep reinforcement learning, a major challenge to widespread industrial adoption of deep reinforcement learning that has recently surfaced but little explored is the potential vulnerability to privacy breaches. In particular, there have been no concrete adversarial attack strategies in literature tailored for studying the vulnerability of deep reinforcement learning algorithms to membership inference attacks. To address this gap, we propose an adversarial attack framework tailored for testing the vulnerability of deep reinforcement learning algorithms to membership inference attacks. More specifically, we design a series of experiments to investigate the impact of temporal correlation, which naturally exists in reinforcement learning training data, on the probability of information leakage. Furthermore, we study the differences in the performance of \emph{collective} and \emph{individual} membership attacks against deep reinforcement learning algorithms. Experimental results show that the proposed adversarial attack framework is surprisingly effective at inferring the data used during deep reinforcement training with an accuracy exceeding $84\%$ in individual and $97\%$ in collective mode on two different control tasks in OpenAI Gym, which raises serious privacy concerns in the deployment of models resulting from deep reinforcement learning. Moreover, we show that the learning state of a reinforcement learning algorithm significantly influences the level of the privacy breach.

(0 min) Building open-domain conversational systems (or chatbots) that produce convincing responses is a recognized challenge. Recent state-of-the-art (SoTA) transformer-based models for the generation of natural language dialogue have demonstrated impressive performance in simulating human-like, single-turn conversations in English. This work investigates, by an empirical study, the potential for transfer learning of such models to Swedish language. DialoGPT, an English language pre-trained model, is adapted by training on three different Swedish language conversational datasets obtained from publicly available sources. Perplexity score (an automated intrinsic language model metric) and surveys by human evaluation were used to assess the performances of the fine-tuned models, with results that indicate that the capacity for transfer learning can be exploited with considerable success. Human evaluators asked to score the simulated dialogue judged over 57% of the chatbot responses to be human-like for the model trained on the largest (Swedish) dataset. We provide the demos and model checkpoints of our English and Swedish chatbots on the HuggingFace platform for public use.

(0 min) Human-AI collaborative policy synthesis is a procedure in which (1) a human initializes an autonomous agent's behavior, (2) Reinforcement Learning improves the human specified behavior, and (3) the agent can explain the final optimized policy to the user. This paradigm leverages human expertise and facilitates a greater insight into the learned behaviors of an agent. Existing approaches to enabling collaborative policy specification involve black box methods which are unintelligible and are not catered towards non-expert end-users. In this paper, we develop a novel collaborative framework to enable humans to initialize and interpret an autonomous agent's behavior, rooted in principles of human-centered design. Through our framework, we enable humans to specify an initial behavior model in the form of unstructured, natural language, which we then convert to lexical decision trees. Next, we are able to leverage these human-specified policies, to warm-start reinforcement learning and further allow the agent to optimize the policies through reinforcement learning. Finally, to close the loop on human-specification, we produce explanations of the final learned policy, in multiple modalities, to provide the user a final depiction about the learned policy of the agent. We validate our approach by showing that our model can produce >80% accuracy, and that human-initialized policies are able to successfully warm-start RL. We then conduct a novel human-subjects study quantifying the relative subjective and objective benefits of varying XAI modalities(e.g., Tree, Language, and Program) for explaining learned policies to end-users, in terms of usability and interpretability and identify the circumstances that influence these measures. Our findings emphasize the need for personalized explainable systems that can facilitate user-centric policy explanations for a variety of end-users.

(0 min) Mainstream approaches for unsupervised domain adaptation (UDA) learn domain-invariant representations to narrow the domain shift. Recently, self-training has been gaining momentum in UDA, which exploits unlabeled target data by training with target pseudo-labels. However, as corroborated in this work, under distributional shift in UDA, the pseudo-labels can be unreliable in terms of their large discrepancy from target ground truth. Thereby, we propose Cycle Self-Training (CST), a principled self-training algorithm that explicitly enforces pseudo-labels to generalize across domains. CST cycles between a forward step and a reverse step until convergence. In the forward step, CST generates target pseudo-labels with a source-trained classifier. In the reverse step, CST trains a target classifier using target pseudo-labels, and then updates the shared representations to make the target classifier perform well on the source data. We introduce the Tsallis entropy as a confidence-friendly regularization to improve the quality of target pseudo-labels. We analyze CST theoretically under realistic assumptions, and provide hard cases where CST recovers target ground truth, while both invariant feature learning and vanilla self-training fail. Empirical results indicate that CST significantly improves over the state-of-the-arts on visual recognition and sentiment analysis benchmarks.

(0 min) While wav2vec 2.0 has been proposed for speech recognition (ASR), it can also be used for speech emotion recognition (SER); its performance can be significantly improved using different fine-tuning strategies. Two baseline methods, vanilla fine-tuning (V-FT) and task adaptive pretraining (TAPT) are first presented. We show that V-FT is able to outperform state-of-the-art models on the IEMOCAP dataset. TAPT, an existing NLP fine-tuning strategy, further improves the performance on SER. We also introduce a novel fine-tuning method termed P-TAPT, which modifies the TAPT objective to learn contextualized emotion representations. Experiments show that P-TAPT performs better than TAPT especially under low-resource settings. Compared to prior works in this literature, our top-line system achieved a 7.4% absolute improvement on unweighted accuracy (UA) over the state-of-the-art performance on IEMOCAP. Our code is publicly available.

(0 min) Improving the quality of Natural Language Understanding (NLU) models, and more specifically, task-oriented semantic parsing models, in production is a cumbersome task. In this work, we present a system called AutoNLU, which we designed to scale the NLU quality improvement process. It adds automation to three key steps: detection, attribution, and correction of model errors, i.e., bugs. We detected four times more failed tasks than with random sampling, finding that even a simple active learning sampling method on an uncalibrated model is surprisingly effective for this purpose. The AutoNLU tool empowered linguists to fix ten times more semantic parsing bugs than with prior manual processes, auto-correcting 65% of all identified bugs.

(0 min) A fundamental bottleneck in utilising complex machine learning systems for critical applications has been not knowing why they do and what they do, thus preventing the development of any crucial safety protocols. To date, no method exist that can provide full insight into the granularity of the neural network's decision process. In the past, saliency maps were an early attempt at resolving this problem through sensitivity calculations, whereby dimensions of a data point are selected based on how sensitive the output of the system is to them. However, the success of saliency maps has been at best limited, mainly due to the fact that they interpret the underlying learning system through a linear approximation. We present a novel class of methods for generating nonlinear saliency maps which fully account for the nonlinearity of the underlying learning system. While agreeing with linear saliency maps on simple problems where linear saliency maps are correct, they clearly identify more specific drivers of classification on complex examples where nonlinearities are more pronounced. This new class of methods significantly aids interpretability of deep neural networks and related machine learning systems. Crucially, they provide a starting point for their more broad use in serious applications, where 'why' is equally important as 'what'.

(0 min) We study the model-based reward-free reinforcement learning with linear function approximation for episodic Markov decision processes (MDPs). In this setting, the agent works in two phases. In the exploration phase, the agent interacts with the environment and collects samples without the reward. In the planning phase, the agent is given a specific reward function and uses samples collected from the exploration phase to learn a good policy. We propose a new provably efficient algorithm, called UCRL-RFE under the Linear Mixture MDP assumption, where the transition probability kernel of the MDP can be parameterized by a linear function over certain feature mappings defined on the triplet of state, action, and next state. We show that to obtain an $\epsilon$-optimal policy for arbitrary reward function, UCRL-RFE needs to sample at most $\tilde O(H^5d^2\epsilon^{-2})$ episodes during the exploration phase. Here, $H$ is the length of the episode, $d$ is the dimension of the feature mapping. We also propose a variant of UCRL-RFE using Bernstein-type bonus and show that it needs to sample at most $\tilde O(H^4d(H + d)\epsilon^{-2})$ to achieve an $\epsilon$-optimal policy. By constructing a special class of linear Mixture MDPs, we also prove that for any reward-free algorithm, it needs to sample at least $\tilde \Omega(H^2d\epsilon^{-2})$ episodes to obtain an $\epsilon$-optimal policy. Our upper bound matches the lower bound in terms of the dependence on $\epsilon$ and the dependence on $d$ if $H \ge d$.

(0 min) Numerous challenges in science and engineering can be framed as optimization tasks, including the maximization of reaction yields, the optimization of molecular and materials properties, and the fine-tuning of automated hardware protocols. Design of experiment and optimization algorithms are often adopted to solve these tasks efficiently. Increasingly, these experiment planning strategies are coupled with automated hardware to enable autonomous experimental platforms. The vast majority of the strategies used, however, do not consider robustness against the variability of experiment and process conditions. In fact, it is generally assumed that these parameters are exact and reproducible. Yet some experiments may have considerable noise associated with some of their conditions, and process parameters optimized under precise control may be applied in the future under variable operating conditions. In either scenario, the optimal solutions found might not be robust against input variability, affecting the reproducibility of results and returning suboptimal performance in practice. Here, we introduce Golem, an algorithm that is agnostic to the choice of experiment planning strategy and that enables robust experiment and process optimization. Golem identifies optimal solutions that are robust to input uncertainty, thus ensuring the reproducible performance of optimized experimental protocols and processes. It can be used to analyze the robustness of past experiments, or to guide experiment planning algorithms toward robust solutions on the fly. We assess the performance and domain of applicability of Golem through extensive benchmark studies and demonstrate its practical relevance by optimizing an analytical chemistry protocol under the presence of significant noise in its experimental conditions.

(0 min) Many time series can be modeled as a sequence of segments representing high-level discrete states, such as running and walking in a human activity application. Flexible models should describe the system state and observations in stationary ``pure-state'' periods as well as transition periods between adjacent segments, such as a gradual slowdown between running and walking. However, most prior work assumes instantaneous transitions between pure discrete states. We propose a dynamical Wasserstein barycentric (DWB) model that estimates the system state over time as well as the data-generating distributions of pure states in an unsupervised manner. Our model assumes each pure state generates data from a multivariate normal distribution, and characterizes transitions between states via displacement-interpolation specified by the Wasserstein barycenter. The system state is represented by a barycentric weight vector which evolves over time via a random walk on the simplex. Parameter learning leverages the natural Riemannian geometry of Gaussian distributions under the Wasserstein distance, which leads to improved convergence speeds. Experiments on several human activity datasets show that our proposed DWB model accurately learns the generating distribution of pure states while improving state estimation for transition periods compared to the commonly used linear interpolation mixture models.

(0 min) The presentation and analysis of image data from a single viewpoint are often not sufficient to solve a task. Several viewpoints are necessary to obtain more information. The $\textit{next-best-view}$ problem attempts to find the optimal viewpoint with the greatest information gain for the underlying task. In this work, a robot arm holds an object in its end-effector and searches for a sequence of next-best-view to explicitly identify the object. We use Soft Actor-Critic (SAC), a method of deep reinforcement learning, to learn these next-best-views for a specific set of objects. The evaluation shows that an agent can learn to determine an object pose to which the robot arm should move an object. This leads to a viewpoint that provides a more accurate prediction to distinguish such an object from other objects better. We make the code publicly available for the scientific community and for reproducibility under $\href{https://github.com/ckorbach/nbv_rl}{\text{this https link}}$.

(0 min) Semi-supervised learning (SSL) has demonstrated its potential to improve the model accuracy for a variety of learning tasks when the high-quality supervised data is severely limited. Although it is often established that the average accuracy for the entire population of data is improved, it is unclear how SSL fares with different sub-populations. Understanding the above question has substantial fairness implications when these different sub-populations are defined by the demographic groups we aim to treat fairly. In this paper, we reveal the disparate impacts of deploying SSL: the sub-population who has a higher baseline accuracy without using SSL (the ``rich" sub-population) tends to benefit more from SSL; while the sub-population who suffers from a low baseline accuracy (the ``poor" sub-population) might even observe a performance drop after adding the SSL module. We theoretically and empirically establish the above observation for a broad family of SSL algorithms, which either explicitly or implicitly use an auxiliary ``pseudo-label". Our experiments on a set of image and text classification tasks confirm our claims. We discuss how this disparate impact can be mitigated and hope that our paper will alarm the potential pitfall of using SSL and encourage a multifaceted evaluation of future SSL algorithms. Code is available at github.com/UCSC-REAL/Disparate-SSL.

(0 min) While deep reinforcement learning (RL) promises freedom from hand-labeled data, great successes, especially for Embodied AI, require significant work to create supervision via carefully shaped rewards. Indeed, without shaped rewards, i.e., with only terminal rewards, present-day Embodied AI results degrade significantly across Embodied AI problems from single-agent Habitat-based PointGoal Navigation (SPL drops from 55 to 0) and two-agent AI2-THOR-based Furniture Moving (success drops from 58% to 1%) to three-agent Google Football-based 3 vs. 1 with Keeper (game score drops from 0.6 to 0.1). As training from shaped rewards doesn't scale to more realistic tasks, the community needs to improve the success of training with terminal rewards. For this we propose GridToPix: 1) train agents with terminal rewards in gridworlds that generically mirror Embodied AI environments, i.e., they are independent of the task; 2) distill the learned policy into agents that reside in complex visual worlds. Despite learning from only terminal rewards with identical models and RL algorithms, GridToPix significantly improves results across tasks: from PointGoal Navigation (SPL improves from 0 to 64) and Furniture Moving (success improves from 1% to 25%) to football gameplay (game score improves from 0.1 to 0.6). GridToPix even helps to improve the results of shaped reward training.

(2 min) We study contextual search, a generalization of binary search in higher dimensions, which captures settings such as feature-based dynamic pricing. Standard game-theoretic formulations of this problem assume that agents act in accordance with a specific behavioral model. In practice, however, some agents may not subscribe to the dominant behavioral model or may act in ways that seem to be arbitrarily irrational. Existing algorithms heavily depend on the behavioral model being (approximately) accurate for all agents and have poor performance in the presence of even a few such arbitrarily irrational agents. We initiate the study of contextual search when some of the agents can behave in ways inconsistent with the underlying behavioral model. In particular, we provide two algorithms, one based on multidimensional binary search methods and one based on gradient descent. We show that these algorithms attain near-optimal regret guarantees in the absence of irrational agents and their performance degrades gracefully with the number of such agents, providing the first results for contextual search in any adversarial noise model. Our techniques draw inspiration from learning theory, game theory, high-dimensional geometry, and convex analysis.

(2 min) The proliferation of Internet-of-Things (IoT) devices and cloud-computing applications over siloed data centers is motivating renewed interest in the collaborative training of a shared model by multiple individual clients via federated learning (FL). To improve the communication efficiency of FL implementations in wireless systems, recent works have proposed compression and dimension reduction mechanisms, along with digital and analog transmission schemes that account for channel noise, fading, and interference. The prior art has mainly focused on star topologies consisting of distributed clients and a central server. In contrast, this paper studies FL over wireless device-to-device (D2D) networks by providing theoretical insights into the performance of digital and analog implementations of decentralized stochastic gradient descent (DSGD). First, we introduce generic digital and analog wireless implementations of communication-efficient DSGD algorithms, leveraging random linear coding (RLC) for compression and over-the-air computation (AirComp) for simultaneous analog transmissions. Next, under the assumptions of convexity and connectivity, we provide convergence bounds for both implementations. The results demonstrate the dependence of the optimality gap on the connectivity and on the signal-to-noise ratio (SNR) levels in the network. The analysis is corroborated by experiments on an image-classification task.

(2 min) The conventional wisdom behind learning deep classification models is to focus on bad-classified examples and ignore well-classified examples that are far from the decision boundary. For instance, when training with cross-entropy loss, examples with higher likelihoods (i.e., well-classified examples) contribute smaller gradients in back-propagation. However, we theoretically show that this common practice hinders representation learning, energy optimization, and the growth of margin. To counteract this deficiency, we propose to reward well-classified examples with additive bonuses to revive their contribution to learning. This counterexample theoretically addresses these three issues. We empirically support this claim by directly verify the theoretical results or through the significant performance improvement with our counterexample on diverse tasks, including image classification, graph classification, and machine translation. Furthermore, this paper shows that because our idea can solve these three issues, we can deal with complex scenarios, such as imbalanced classification, OOD detection, and applications under adversarial attacks.

(2 min) Robust Markov decision processes (MDPs) aim to handle changing or partially known system dynamics. To solve them, one typically resorts to robust optimization methods. However, this significantly increases computational complexity and limits scalability in both learning and planning. On the other hand, regularized MDPs show more stability in policy learning without impairing time complexity. Yet, they generally do not encompass uncertainty in the model dynamics. In this work, we aim to learn robust MDPs using regularization. We first show that regularized MDPs are a particular instance of robust MDPs with uncertain reward. We thus establish that policy iteration on reward-robust MDPs can have the same time complexity as on regularized MDPs. We further extend this relationship to MDPs with uncertain transitions: this leads to a regularization term with an additional dependence on the value function. We finally generalize regularized MDPs to twice regularized MDPs (R${}^2$ MDPs), i.e., MDPs with $\textit{both}$ value and policy regularization. The corresponding Bellman operators enable developing policy iteration schemes with convergence and robustness guarantees. It also reduces planning and learning in robust MDPs to regularized MDPs.

(2 min) In recent years, Artificial Intelligence (AI) algorithms have been proven to outperform traditional statistical methods in terms of predictivity, especially when a large amount of data was available. Nevertheless, the "black box" nature of AI models is often a limit for a reliable application in high-stakes fields like diagnostic techniques, autonomous guide, etc. Recent works have shown that an adequate level of interpretability could enforce the more general concept of model trustworthiness. The basic idea of this paper is to exploit the human prior knowledge of the features' importance for a specific task, in order to coherently aid the phase of the model's fitting. This sort of "weighted" AI is obtained by extending the empirical loss with a regularization term encouraging the importance of the features to follow predetermined constraints. This procedure relies on local methods for the feature importance computation, e.g. LRP, LIME, etc. that are the link between the model weights to be optimized and the user-defined constraints on feature importance. In the fairness area, promising experimental results have been obtained for the Adult dataset. Many other possible applications of this model agnostic theoretical framework are described.

(2 min) The early detection of terrorist threat objects, such as guns and knives, through improved metal detection, has the potential to reduce the number of attacks and improve public safety and security. To achieve this, there is considerable potential to use the fields applied and measured by a metal detector to discriminate between different shapes and different metals since, hidden within the field perturbation, is object characterisation information. The magnetic polarizability tensor (MPT) offers an economical characterisation of metallic objects and its spectral signature provides additional object characterisation information. The MPT spectral signature can be determined from measurements of the induced voltage over a range frequencies in a metal signature for a hidden object. With classification in mind, it can also be computed in advance for different threat and non-threat objects. In the article, we evaluate the performance of probabilistic and non-probabilistic machine learning algorithms, trained using a dictionary of computed MPT spectral signatures, to classify objects for metal detection. We discuss the importances of using appropriate features and selecting an appropriate algorithm depending on the classification problem being solved and we present numerical results for a range of practically motivated metal detection classification problems.

(2 min) Recent years have witnessed the dramatic growth of paper volumes with plenty of new research papers published every day, especially in the area of computer science. How to glean papers worth reading from the massive literature to do a quick survey or keep up with the latest advancement about a specific research topic has become a challenging task. Existing academic search engines such as Google Scholar return relevant papers by individually calculating the relevance between each paper and query. However, such systems usually omit the prerequisite chains of a research topic and cannot form a meaningful reading path. In this paper, we introduce a new task named Reading Path Generation (RPG) which aims at automatically producing a path of papers to read for a given query. To serve as a research benchmark, we further propose SurveyBank, a dataset consisting of large quantities of survey papers in the field of computer science as well as their citation relationships. Each survey paper contains key phrases extracted from its title and multi-level reading lists inferred from its references. Furthermore, we propose a graph-optimization-based approach for reading path generation which takes the relationship between papers into account. Extensive evaluations demonstrate that our approach outperforms other baselines. A Real-time Reading Path Generation System (RePaGer) has been also implemented with our designed model. To the best of our knowledge, we are the first to target this important research problem. Our source code of RePaGer system and SurveyBank dataset can be found on here.

(2 min) We consider off-policy evaluation (OPE) in continuous treatment settings, such as personalized dose-finding. In OPE, one aims to estimate the mean outcome under a new treatment decision rule using historical data generated by a different decision rule. Most existing works on OPE focus on discrete treatment settings. To handle continuous treatments, we develop a novel estimation method for OPE using deep jump learning. The key ingredient of our method lies in adaptively discretizing the treatment space using deep discretization, by leveraging deep learning and multi-scale change point detection. This allows us to apply existing OPE methods in discrete treatments to handle continuous treatments. Our method is further justified by theoretical results, simulations, and a real application to Warfarin Dosing.

(2 min) The separation of the data capture and analysis in modern vision systems has led to a massive amount of data transfer between the end devices and cloud computers, resulting in long latency, slow response, and high power consumption. Efficient hardware architectures are under focused development to enable Artificial Intelligence (AI) at the resource-limited end sensing devices. One of the most promising solutions is to enable Processing-in-Pixel (PIP) scheme. However, the conventional schemes suffer from the low fill-factor issue. This paper proposes a PIP based CMOS sensor architecture, which allows convolution operation before the column readout circuit to significantly improve the image reading speed with much lower power consumption. The simulation results show that the proposed architecture could support the computing efficiency up to 11.65 TOPS/W at the 8-bit weight configuration, which is three times as high as the conventional schemes. The transistors required for each pixel are only 2.5T, significantly improving the fill-factor.

(2 min) Recent work has argued that classification losses utilizing softmax cross-entropy are superior not only for fixed-set classification tasks, but also by outperforming losses developed specifically for open-set tasks including few-shot learning and retrieval. Softmax classifiers have been studied using different embedding geometries -- Euclidean, hyperbolic, and spherical -- and claims have been made about the superiority of one or another, but they have not been systematically compared with careful controls. We conduct an empirical investigation of embedding geometry on softmax losses for a variety of fixed-set classification and image retrieval tasks. An interesting property observed for the spherical losses lead us to propose a probabilistic classifier based on the von Mises-Fisher distribution, and we show that it is competitive with state-of-the-art methods while producing improved out-of-the-box calibration. We provide guidance regarding the trade-offs between losses and how to choose among them.

(0 min) Understanding the implicit bias of Stochastic Gradient Descent (SGD) is one of the key challenges in deep learning, especially for overparametrized models, where the local minimizers of the loss function $L$ can form a manifold. Intuitively, with a sufficiently small learning rate $\eta$, SGD tracks Gradient Descent (GD) until it gets close to such manifold, where the gradient noise prevents further convergence. In such a regime, Blanc et al. (2020) proved that SGD with label noise locally decreases a regularizer-like term, the sharpness of loss, $\mathrm{tr}[\nabla^2 L]$. The current paper gives a general framework for such analysis by adapting ideas from Katzenberger (1991). It allows in principle a complete characterization for the regularization effect of SGD around such manifold -- i.e., the "implicit bias" -- using a stochastic differential equation (SDE) describing the limiting dynamics of the parameters, which is determined jointly by the loss function and the noise covariance. This yields some new results: (1) a global analysis of the implicit bias valid for $\eta^{-2}$ steps, in contrast to the local analysis of Blanc et al. (2020) that is only valid for $\eta^{-1.6}$ steps and (2) allowing arbitrary noise covariance. As an application, we show with arbitrary large initialization, label noise SGD can always escape the kernel regime and only requires $O(\kappa\ln d)$ samples for learning an $\kappa$-sparse overparametrized linear model in $\mathbb{R}^d$ (Woodworth et al., 2020), while GD initialized in the kernel regime requires $\Omega(d)$ samples. This upper bound is minimax optimal and improves the previous $\tilde{O}(\kappa^2)$ upper bound (HaoChen et al., 2020).

(0 min) Emphatic temporal difference (ETD) learning (Sutton et al., 2016) is a successful method to conduct the off-policy value function evaluation with function approximation. Although ETD has been shown to converge asymptotically to a desirable value function, it is well-known that ETD often encounters a large variance so that its sample complexity can increase exponentially fast with the number of iterations. In this work, we propose a new ETD method, called PER-ETD (i.e., PEriodically Restarted-ETD), which restarts and updates the follow-on trace only for a finite period for each iteration of the evaluation parameter. Further, PER-ETD features a design of the logarithmical increase of the restart period with the number of iterations, which guarantees the best trade-off between the variance and bias and keeps both vanishing sublinearly. We show that PER-ETD converges to the same desirable fixed point as ETD, but improves the exponential sample complexity of ETD to be polynomials. Our experiments validate the superior performance of PER-ETD and its advantage over ETD.

(0 min) Vibration-based techniques are among the most common condition monitoring approaches. With the advancement of computers, these approaches have also been improved such that recently, these approaches in conjunction with deep learning methods attract attention among researchers. This is mostly due to the nature of the deep learning method that could facilitate the monitoring procedure by integrating the feature extraction, feature selection, and classification steps into one automated step. However, this can be achieved at the expense of challenges in designing the architecture of a deep learner, tuning its hyper-parameters. Moreover, it sometimes gives low generalization capability. As a remedy to these problems, this study proposes a framework based on ensemble deep learning methodology. The framework was initiated by creating a pool of Convolutional neural networks (CNN). To create diversity to the CNNs, they are fed by frequency responses which are passed through different functions. As the next step, proper CNNs are selected based on an information criterion to be used for fusion. The fusion is then carried out by improved Dempster-Shafer theory. The proposed framework is applied to real test data collected from Equiax Polycrystalline Nickel alloy first-stage turbine blades with complex geometry.

(0 min) Although conceptualization has been widely studied in semantics and knowledge representation, it is still challenging to find the most accurate concept phrases to characterize the main idea of a text snippet on the fast-growing social media. This is partly attributed to the fact that most knowledge bases contain general terms of the world, such as trees and cars, which do not have the defining power or are not interesting enough to social media app users. Another reason is that the intricacy of natural language allows the use of tense, negation and grammar to change the logic or emphasis of language, thus conveying completely different meanings. In this paper, we present TAG, a high-quality concept matching dataset consisting of 10,000 labeled pairs of fine-grained concepts and web-styled natural language sentences, mined from the open-domain social media. The concepts we consider represent the trending interests of online users. Associated with TAG is a concept graph of these fine-grained concepts and entities to provide the structural context information. We evaluate a wide range of popular neural text matching models as well as pre-trained language models on TAG, and point out their insufficiency to tag social media content with the most appropriate concept. We further propose a novel graph-graph matching method that demonstrates superior abstraction and generalization performance by better utilizing both the structural context in the concept graph and logic interactions between semantic units in the sentence via syntactic dependency parsing. We open-source both the TAG dataset and the proposed methods to facilitate further research.

(2 min) While Spiking Neural Networks (SNNs) have been gaining in popularity, it seems that the algorithms used to train them are not powerful enough to solve the same tasks as those tackled by classical Artificial Neural Networks (ANNs). In this paper, we provide an extra tool to help us understand and train SNNs by using theory from the field of time encoding. Time encoding machines (TEMs) can be used to model integrate-and-fire neurons and have well-understood reconstruction properties. We will see how one can take inspiration from the field of TEMs to interpret the spike times of SNNs as constraints on the SNNs' weight matrices. More specifically, we study how to train one-layer SNNs by solving a set of linear constraints, and how to train two-layer SNNs by leveraging the all-or-none and asynchronous properties of the spikes emitted by SNNs. These properties of spikes result in an alternative to backpropagation which is not possible in the case of simultaneous and graded activations as in classical ANNs.

(2 min) Quantification is the research field that studies methods for counting the number of data points that belong to each class in an unlabeled sample. Traditionally, researchers in this field assume the availability of labelled observations for all classes to induce a quantification model. However, we often face situations where the number of classes is large or even unknown, or we have reliable data for a single class. When inducing a multi-class quantifier is infeasible, we are often concerned with estimates for a specific class of interest. In this context, we have proposed a novel setting known as One-class Quantification (OCQ). In contrast, Positive and Unlabeled Learning (PUL), another branch of Machine Learning, has offered solutions to OCQ, despite quantification not being the focal point of PUL. This article closes the gap between PUL and OCQ and brings both areas together under a unified view. We compare our method, Passive Aggressive Threshold (PAT), against PUL methods and show that PAT generally is the fastest and most accurate algorithm. PAT induces quantification models that can be reused to quantify different samples of data. We additionally introduce Exhaustive TIcE (ExTIcE), an improved version of the PUL algorithm Tree Induction for c Estimation (TIcE). We show that ExTIcE quantifies more accurately than PAT and the other assessed algorithms in scenarios where several negative observations are identical to the positive ones.

(2 min) We study generalization properties of random features (RF) regression in high dimensions optimized by stochastic gradient descent (SGD). In this regime, we derive precise non-asymptotic error bounds of RF regression under both constant and adaptive step-size SGD setting, and observe the double descent phenomenon both theoretically and empirically. Our analysis shows how to cope with multiple randomness sources of initialization, label noise, and data sampling (as well as stochastic gradients) with no closed-form solution, and also goes beyond the commonly-used Gaussian/spherical data assumption. Our theoretical results demonstrate that, with SGD training, RF regression still generalizes well for interpolation learning, and is able to characterize the double descent behavior by the unimodality of variance and monotonic decrease of bias. Besides, we also prove that the constant step-size SGD setting incurs no loss in convergence rate when compared to the exact minimal-norm interpolator, as a theoretical justification of using SGD in practice.

(0 min) There is a growing demand for redirected walking (RDW) techniques and their application. To apply appropriate RDW methods and manipulation, the RDW controllers are predominantly used. There are three types of RDW controllers: direct scripted controller, generalized controller, and predictive controller. The scripted controller type pre-scripts the mapping between the real and virtual environments. The generalized controller type employs the RDW method and manipulation quantities according to a certain procedure depending on the user's position in relation to the real space. This approach has the potential to be reused in any environment; however, it is not fully optimized. The predictive controller type predicts the user's future path using the user's behavior and manages RDW techniques. This approach is highly anticipated to be very effective and versatile; however, it has not been sufficiently developed. This paper proposes a novel RDW controller using reinforcement learning (RL) with advanced plannability/versatility. Our simulation experiments indicate that the proposed method can reduce the number of reset manipulations, which is one of the indicators of the effectiveness of the RDW controller, compared to the generalized controller under real environments with many obstacles. Meanwhile, the experimental results also showed that the gain output by the proposed method oscillates. The results of a user study conducted showed that the proposed RDW controller can reduce the number of resets compared to the conventional generalized controller. Furthermore, no adverse effects such as cybersickness associated with the oscillation of the output gain were evinced. The simulation and user studies demonstrate that the proposed RDW controller with RL outperforms the existing generalized controllers and can be applied to users.

(0 min) The paper considers a Mixture Multilayer Stochastic Block Model (MMLSBM), where layers can be partitioned into groups of similar networks, and networks in each group are equipped with a distinct Stochastic Block Model. The goal is to partition the multilayer network into clusters of similar layers, and to identify communities in those layers. Jing et al. (2020) introduced the MMLSBM and developed a clustering methodology, TWIST, based on regularized tensor decomposition. The present paper proposes a different technique, an alternating minimization algorithm (ALMA), that aims at simultaneous recovery of the layer partition, together with estimation of the matrices of connection probabilities of the distinct layers. Compared to TWIST, ALMA achieves higher accuracy both theoretically and numerically.

(0 min) Activity recognition systems that are capable of estimating human activities from wearable inertial sensors have come a long way in the past decades. Not only have state-of-the-art methods moved away from feature engineering and have fully adopted end-to-end deep learning approaches, best practices for setting up experiments, preparing datasets, and validating activity recognition approaches have similarly evolved. This tutorial was first held at the 2021 ACM International Symposium on Wearable Computers (ISWC'21) and International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp'21). The tutorial, after a short introduction in the research field of activity recognition, provides a hands-on and interactive walk-through of the most important steps in the data pipeline for the deep learning of human activities. All presentation slides shown during the tutorial, which also contain links to all code exercises, as well as the link of the GitHub page of the tutorial can be found on: https://mariusbock.github.io/dl-for-har

(0 min) Within the field of process mining, several different trace clustering approaches exist for partitioning traces or process instances into similar groups. Typically, this partitioning is based on certain patterns or similarity between the traces, or driven by the discovery of a process model for each cluster. The main drawback of these techniques, however, is that their solutions are usually hard to evaluate or justify by domain experts. In this paper, we present two constrained trace clustering techniques that are capable to leverage expert knowledge in the form of instance-level constraints. In an extensive experimental evaluation using two real-life datasets, we show that our novel techniques are indeed capable of producing clustering solutions that are more justifiable without a substantial negative impact on their quality.

(0 min) Federated learning (FL) literature typically assumes that each client has a fixed amount of data, which is unrealistic in many practical applications. Some recent works introduced a framework for online FL (Online-Fed) wherein clients perform model learning on streaming data and communicate the model to the server; however, they do not address the associated communication overhead. As a solution, this paper presents a partial-sharing-based online federated learning framework (PSO-Fed) that enables clients to update their local models using continuous streaming data and share only portions of those updated models with the server. During a global iteration of PSO-Fed, non-participant clients have the privilege to update their local models with new data. Here, we consider a global task of kernel regression, where clients use a random Fourier features-based kernel LMS on their data for local learning. We examine the mean convergence of the PSO-Fed for kernel regression. Experimental results show that PSO-Fed can achieve competitive performance with a significantly lower communication overhead than Online-Fed.

(0 min) The FJet approach is introduced for determining the underlying model of a dynamical system. It borrows ideas from the fields of Lie symmetries as applied to differential equations (DEs), and numerical integration (such as Runge-Kutta). The technique can be considered as a way to use machine learning (ML) to derive a numerical integration scheme. The technique naturally overcomes the "extrapolation problem", which is when ML is used to extrapolate a model beyond the time range of the original training data. It does this by doing the modeling in the phase space of the system, rather than over the time domain. When modeled with a type of regression scheme, it's possible to accurately determine the underlying DE, along with parameter dependencies. Ideas from the field of Lie symmetries applied to ordinary DEs are used to determine constants of motion, even for damped and driven systems. These statements are demonstrated on three examples: a damped harmonic oscillator, a damped pendulum, and a damped, driven nonlinear oscillator (Duffing oscillator). In the model for the Duffing oscillator, it's possible to treat the external force in a manner reminiscent of a Green's function approach. Also, in the case of the undamped harmonic oscillator, the FJet approach remains stable approximately $10^9$ times longer than $4$th-order Runge-Kutta.

(0 min) In this paper, we conjecture that if the permutation invariance of neural networks is taken into account, SGD solutions will likely have no barrier in the linear interpolation between them. Although it is a bold conjecture, we show how extensive empirical attempts fall short of refuting it. We further provide a preliminary theoretical result to support our conjecture. Our conjecture has implications for lottery ticket hypothesis, distributed training, and ensemble methods.

(0 min) The goal in a blind image quality assessment (BIQA) model is to simulate the process of evaluating images by human eyes and accurately assess the quality of the image. Although many approaches effectively identify degradation, they do not fully consider the semantic content in images resulting in distortion. In order to fill this gap, we propose a deep adaptive superpixel-based network, namely DSN-IQA, to assess the quality of image based on multi-scale and superpixel segmentation. The DSN-IQA can adaptively accept arbitrary scale images as input images, making the assessment process similar to human perception. The network uses two models to extract multi-scale semantic features and generate a superpixel adjacency map. These two elements are united together via feature fusion to accurately predict image quality. Experimental results on different benchmark databases demonstrate that our algorithm is highly competitive with other approaches when assessing challenging authentic image databases. Also, due to adaptive deep superpixel-based network, our model accurately assesses images with complicated distortion, much like the human eye.

(0 min) A deep generative model is characterized by a representation space, its distribution, and a neural network mapping the representation to a distribution over vectors in feature space. Common methods such as variational autoencoders (VAEs) apply variational inference for training the neural network, but optimizing these models is often non-trivial. The encoder adds to the complexity of the model and introduces an amortization gap and the quality of the variational approximation is usually unknown. Additionally, the balance of the loss terms of the objective function heavily influences performance. Therefore, we argue that it is worthwhile to investigate a much simpler approximation which finds representations and their distribution by maximizing the model likelihood via back-propagation. In this approach, there is no encoder, and we therefore call it a Deep Generative Decoder (DGD). Using the CIFAR10 data set, we show that the DGD is easier and faster to optimize than the VAE, achieves more consistent low reconstruction errors of test data, and alleviates the problem of balancing the reconstruction and distribution loss terms. Although the model in its simple form cannot compete with state-of-the-art image generation approaches, it obtains better image generation scores than the variational approach on the CIFAR10 data. We demonstrate on MNIST data how the use of a Gaussian mixture with priors can lead to a clear separation of classes in a 2D representation space, and how the DGD can be used with labels to obtain a supervised representation.

(0 min) Pairwise dot product-based attention allows Transformers to exchange information between tokens in an input-dependent way, and is key to their success across diverse applications in language and vision. However, a typical Transformer model computes such pairwise attention scores repeatedly for the same sequence, in multiple heads in multiple layers. We systematically analyze the empirical similarity of these scores across heads and layers and find them to be considerably redundant, especially adjacent layers showing high similarity. Motivated by these findings, we propose a novel architecture that reuses attention scores computed in one layer in multiple subsequent layers. Experiments on a number of standard benchmarks show that reusing attention delivers performance equivalent to or better than standard transformers, while reducing both compute and memory usage.

(0 min) Neurons in the brain are complex machines with distinct functional compartments that interact nonlinearly. In contrast, neurons in artificial neural networks abstract away this complexity, typically down to a scalar activation function of a weighted sum of inputs. Here we emulate more biologically realistic neurons by learning canonical activation functions with two input arguments, analogous to basal and apical dendrites. We use a network-in-network architecture where each neuron is modeled as a multilayer perceptron with two inputs and a single output. This inner perceptron is shared by all units in the outer network. Remarkably, the resultant nonlinearities reliably produce soft XOR functions, consistent with recent experimental observations about interactions between inputs in human cortical neurons. When hyperparameters are optimized, networks with these nonlinearities learn faster and perform better than conventional ReLU nonlinearities with matched parameter counts, and they are more robust to natural and adversarial perturbations.

(0 min) Deep Learning (DL) requires a large amount of training data to provide quality outcomes. However, the field of medical imaging suffers from the lack of sufficient data for properly training DL models because medical images require manual labelling carried out by clinical experts thus the process is time-consuming, expensive, and error-prone. Recently, transfer learning (TL) was introduced to reduce the need for the annotation procedure by means of transferring the knowledge performed by a previous task and then fine-tuning the result using a relatively small dataset. Nowadays, multiple classification methods from medical imaging make use of TL from general-purpose pre-trained models, e.g., ImageNet, which has been proven to be ineffective due to the mismatch between the features learned from natural images (ImageNet) and those more specific from medical images especially medical gray images such as X-rays. ImageNet does not have grayscale images such as MRI, CT, and X-ray. In this paper, we propose a novel DL model to be used for addressing classification tasks of medical imaging, called MedNet. To do so, we aim to issue two versions of MedNet. The first one is Gray-MedNet which will be trained on 3M publicly available gray-scale medical images including MRI, CT, X-ray, ultrasound, and PET. The second version is Color-MedNet which will be trained on 3M publicly available color medical images including histopathology, taken images, and many others. To validate the effectiveness MedNet, both versions will be fine-tuned to train on the target tasks of a more reduced set of medical images. MedNet performs as the pre-trained model to tackle any real-world application from medical imaging and achieve the level of generalization needed for dealing with medical imaging tasks, e.g. classification. MedNet would serve the research community as a baseline for future research.

(0 min) Modern pattern recognition tasks use complex algorithms that take advantage of large datasets to make more accurate predictions than traditional algorithms such as decision trees or k-nearest-neighbor better suited to describe simple structures. While increased accuracy is often crucial, less complexity also has value. This paper proposes a training data selection algorithm that identifies multiple subsets with simple structures. A learning algorithm trained on such a subset can classify an instance belonging to the subset with better accuracy than the traditional learning algorithms. In other words, while existing pattern recognition algorithms attempt to learn a global mapping function to represent the entire dataset, we argue that an ensemble of simple local patterns may better describe the data. Hence the sub-setting algorithm identifies multiple subsets with simple local patterns by identifying similar instances in the neighborhood of an instance. This motivation has similarities to that of gradient boosted trees but focuses on the explainability of the model that is missing for boosted trees. The proposed algorithm thus balances accuracy and explainable machine learning by identifying a limited number of subsets with simple structures. We applied the proposed algorithm to the international stroke dataset to predict the probability of survival. Our bottom-up sub-setting algorithm performed on an average 15% better than the top-down decision tree learned on the entire dataset. The different decision trees learned on the identified subsets use some of the previously unused features by the whole dataset decision tree, and each subset represents a distinct population of data.

(0 min) Training deep neural networks is a well-known highly non-convex problem. In recent works, it is shown that there is no duality gap for regularized two-layer neural networks with ReLU activation, which enables global optimization via convex programs. For multi-layer linear networks with vector outputs, we formulate convex dual problems and demonstrate that the duality gap is non-zero for depth three and deeper networks. However, by modifying the deep networks to more powerful parallel architectures, we show that the duality gap is exactly zero. Therefore, strong convex duality holds, and hence there exist equivalent convex programs that enable training deep networks to global optimality. We also demonstrate that the weight decay regularization in the parameters explicitly encourages low-rank solutions via closed-form expressions. For three-layer non-parallel ReLU networks, we show that strong duality holds for rank-1 data matrices, however, the duality gap is non-zero for whitened data matrices. Similarly, by transforming the neural network architecture into a corresponding parallel version, the duality gap vanishes.

(0 min) The travel marketing platform of Alibaba serves an indispensable role for hundreds of different travel scenarios from Fliggy, Taobao, Alipay apps, etc. To provide personalized recommendation service for users visiting different scenarios, there are two critical issues to be carefully addressed. First, since the traffic characteristics of different scenarios, it is very challenging to train a unified model to serve all. Second, during the promotion period, the exposure of some specific items will be re-weighted due to manual intervention, resulting in biased logs, which will degrade the ranking model trained using these biased data. In this paper, we propose a novel Scenario-Aware Ranking Network (SAR-Net) to address these issues. SAR-Net harvests the abundant data from different scenarios by learning users' cross-scenario interests via two specific attention modules, which leverage the scenario features and item features to modulate the user behavior features, respectively. Then, taking the encoded features of previous module as input, a scenario-specific linear transformation layer is adopted to further extract scenario-specific features, followed by two groups of debias expert networks, i.e., scenario-specific experts and scenario-shared experts. They output intermediate results independently, which are further fused into the final result by a multi-scenario gating module. In addition, to mitigate the data fairness issue caused by manual intervention, we propose the concept of Fairness Coefficient (FC) to measures the importance of individual sample and use it to reweigh the prediction in the debias expert networks. Experiments on an offline dataset covering over 80 million users and 1.55 million travel items and an online A/B test demonstrate the effectiveness of our SAR-Net and its superiority over state-of-the-art methods.

(0 min) It is well-known DNNs would generate different prediction results even given the same model configuration and training dataset. As a result, it becomes more and more important to study prediction variation, i.e. the variation of the predictions on a given input example, in neural network models. Dropout has been commonly used in various applications to quantify prediction variations. However, using dropout in practice can be expensive as it requires running dropout inference many times to estimate prediction variation. In this paper, we study how to estimate dropout prediction variation in a resource-efficient manner. In particular, we demonstrate that we can use neuron activation strength to estimate dropout prediction variation under different dropout settings and on a variety of tasks using three large datasets, MovieLens, Criteo, and EMNIST. Our approach provides an inference-once alternative to estimate dropout prediction variation as an auxiliary task when the main prediction model is served. Moreover, we show that using activation strength features from a subset of neural network layers can be sufficient to achieve similar variation estimation performance compared to using activation features from all layers. This can provide further resource reduction for variation estimation.

(0 min) Graph neural network (GNN) models have achieved great success on graph representation learning. Challenged by large scale private data collection from user-side, GNN models may not be able to reflect the excellent performance, without rich features and complete adjacent relationships. Addressing to the problem, vertical federated learning (VFL) is proposed to implement local data protection through training a global model collaboratively. Consequently, for graph-structured data, it is natural idea to construct VFL framework with GNN models. However, GNN models are proven to be vulnerable to adversarial attacks. Whether the vulnerability will be brought into the VFL has not been studied. In this paper, we devote to study the security issues of GNN based VFL (GVFL), i.e., robustness against adversarial attacks. Further, we propose an adversarial attack method, named Graph-Fraudster. It generates adversarial perturbations based on the noise-added global node embeddings via GVFL's privacy leakage, and the gradient of pairwise node. First, it steals the global node embeddings and sets up a shadow server model for attack generator. Second, noises are added into node embeddings to confuse the shadow server model. At last, the gradient of pairwise node is used to generate attacks with the guidance of noise-added node embeddings. To the best of our knowledge, this is the first study of adversarial attacks on GVFL. The extensive experiments on five benchmark datasets demonstrate that Graph-Fraudster performs better than three possible baselines in GVFL. Furthermore, Graph-Fraudster can remain a threat to GVFL even if two possible defense mechanisms are applied. This paper reveals that GVFL is vulnerable to adversarial attack similar to centralized GNN models.

(0 min) Researchers have been facing a difficult problem that data generation mechanisms could be influenced by internal or external factors leading to the training and test data with quite different distributions, consequently traditional classification or regression from the training set is unable to achieve satisfying results on test data. In this paper, we address this nontrivial domain generalization problem by finding a central subspace in which domain-based covariance is minimized while the functional relationship is simultaneously maximally preserved. We propose a novel variance measurement for multiple domains so as to minimize the difference between conditional distributions across domains with solid theoretical demonstration and supports, meanwhile, the algorithm preserves the functional relationship via maximizing the variance of conditional expectations given output. Furthermore, we also provide a fast implementation that requires much less computation and smaller memory for large-scale matrix operations, suitable for not only domain generalization but also other kernel-based eigenvalue decompositions. To show the practicality of the proposed method, we compare our methods against some well-known dimension reduction and domain generalization techniques on both synthetic data and real-world applications. We show that for small-scale datasets, we are able to achieve better quantitative results indicating better generalization performance over unseen test datasets. For large-scale problems, the proposed fast implementation maintains the quantitative performance but at a substantially lower computational cost.

(0 min) This paper addresses the problem of fake news detection in Spanish using Machine Learning techniques. It is fundamentally the same problem tackled for the English language; however, there is not a significant amount of publicly available and adequately labeled fake news in Spanish to effectively train a Machine Learning model, similarly to those proposed for the English language. Therefore, this work explores different training strategies and architectures to establish a baseline for further research in this area. Four datasets were used, two in English and two in Spanish, and four experimental schemes were tested, including a baseline with classical Machine Learning models, trained and validated using a small dataset in Spanish. The remaining schemes include state-of-the-art Deep Learning models trained (or fine-tuned) and validated in English, trained and validated in Spanish, and fitted in English and validated with automatic translated Spanish sentences. The Deep Learning architectures were built on top of different pre-trained Word Embedding representations, including GloVe, ELMo, BERT, and BETO (a BERT version trained on a large corpus in Spanish). According to the results, the best strategy was a combination of a pre-trained BETO model and a Recurrent Neural Network based on LSTM layers, yielding an accuracy of up to 80%; nonetheless, a baseline model using a Random Forest estimator obtained similar outcomes. Additionally, the translation strategy did not yield acceptable results because of the propagation error; there was also observed a significant difference in models performance when trained in English or Spanish, mainly attributable to the number of samples available for each language.

(0 min) We consider an extended notion of reinforcement learning in which the environment can simulate the agent and base its outputs on the agent's hypothetical behavior. Since good performance usually requires paying attention to whatever things the environment's outputs are based on, we argue that for an agent to achieve on-average good performance across many such extended environments, it is necessary for the agent to self-reflect. Thus, an agent's self-reflection ability can be numerically estimated by running the agent through a battery of extended environments. We are simultaneously releasing an open-source library of extended environments to serve as proof-of-concept of this technique. As the library is first-of-kind, we have avoided the difficult problem of optimizing it. Instead we have chosen environments with interesting properties. Some seem paradoxical, some lead to interesting thought experiments, some are even suggestive of how self-reflection might have evolved in nature. We give examples and introduce a simple transformation which experimentally seems to increase self-reflection.

(0 min) In this paper, we discuss the initial attempts at boosting understanding human language based on deep-learning models with quantum computing. We successfully train a quantum-enhanced Long Short-Term Memory network to perform the parts-of-speech tagging task via numerical simulations. Moreover, a quantum-enhanced Transformer is proposed to perform the sentiment analysis based on the existing dataset.

(0 min) In this paper, we present a novel architecture to realize fine-grained style control on the transformer-based text-to-speech synthesis (TransformerTTS). Specifically, we model the speaking style by extracting a time sequence of local style tokens (LST) from the reference speech. The existing content encoder in TransformerTTS is then replaced by our designed cross-attention blocks for fusion and alignment between content and style. As the fusion is performed along with the skip connection, our cross-attention block provides a good inductive bias to gradually infuse the phoneme representation with a given style. Additionally, we prevent the style embedding from encoding linguistic content by randomly truncating LST during training and using wav2vec 2.0 features. Experiments show that with fine-grained style control, our system performs better in terms of naturalness, intelligibility, and style transferability. Our code and samples are publicly available.

(0 min) Label noise is pervasive in real-world datasets, which encodes wrong correlation patterns and impairs the generalization of deep neural networks (DNNs). It is critical to find efficient ways to detect the corrupted patterns. Current methods primarily focus on designing robust training techniques to prevent DNNs from memorizing corrupted patterns. This approach has two outstanding caveats: 1) applying this approach to each individual dataset would often require customized training processes; 2) as long as the model is trained with noisy supervisions, overfitting to corrupted patterns is often hard to avoid, leading to performance drop in detection. In this paper, given good representations, we propose a universally applicable and training-free solution to detect noisy labels. Intuitively, good representations help define ``neighbors'' of each training instance, and closer instances are more likely to share the same clean label. Based on the neighborhood information, we propose two methods: the first one uses ``local voting" via checking the noisy label consensuses of nearby representations. The second one is a ranking-based approach that scores each instance and filters out a guaranteed number of instances that are likely to be corrupted, again using only representations. Given good (but possibly imperfect) representations that are commonly available in practice, we theoretically analyze how they affect the local voting and provide guidelines for tuning neighborhood size. We also prove the worst-case error bound for the ranking-based method. Experiments with both synthetic and real-world label noise demonstrate our training-free solutions are consistently and significantly improving over most of the training-based baselines. Code is available at github.com/UCSC-REAL/SimiRep.

(0 min) Learning often involves sensitive data and as such, privacy preserving extensions to Stochastic Gradient Descent (SGD) and other machine learning algorithms have been developed using the definitions of Differential Privacy (DP). In differentially private SGD, the gradients computed at each training iteration are subject to two different types of noise. Firstly, inherent sampling noise arising from the use of minibatches. Secondly, additive Gaussian noise from the underlying mechanisms that introduce privacy. In this study, we show that these two types of noise are equivalent in their effect on the utility of private neural networks, however they are not accounted for equally in the privacy budget. Given this observation, we propose a training paradigm that shifts the proportions of noise towards less inherent and more additive noise, such that more of the overall noise can be accounted for in the privacy budget. With this paradigm, we are able to improve on the state-of-the-art in the privacy/utility tradeoff of private end-to-end CNNs.

(0 min) In this paper, we present a new data-driven method for learning stable models of nonlinear systems. Our model lifts the original state space to a higher-dimensional linear manifold using Koopman embeddings. Interestingly, we prove that every discrete-time nonlinear contracting model can be learnt in our framework. Another significant merit of the proposed approach is that it allows for unconstrained optimization over the Koopman embedding and operator jointly while enforcing stability of the model, via a direct parameterization of stable linear systems, greatly simplifying the computations involved. We validate our method on a simulated system and analyze the advantages of our parameterization compared to alternatives.

(0 min) We consider the problem of using expert data with unobserved confounders for imitation and reinforcement learning. We begin by defining the problem of learning from confounded expert data in a contextual MDP setup. We analyze the limitations of learning from such data with and without external reward, and propose an adjustment of standard imitation learning algorithms to fit this setup. We then discuss the problem of distribution shift between the expert data and the online environment when the data is only partially observable. We prove possibility and impossibility results for imitation learning under arbitrary distribution shift of the missing covariates. When additional external reward is provided, we propose a sampling procedure that addresses the unknown shift and prove convergence to an optimal solution. Finally, we validate our claims empirically on challenging assistive healthcare and recommender system simulation tasks.

(0 min) Machine learning and neural network models in particular have been improving the state of the art performance on many artificial intelligence related tasks. Neural network models are typically implemented using frameworks that perform gradient based optimization methods to fit a model to a dataset. These frameworks use a technique of calculating derivatives called automatic differentiation (AD) which removes the burden of performing derivative calculations from the model designer. In this report we describe AD, its motivations, and different implementation approaches. We briefly describe dataflow programming as it relates to AD. Lastly, we present example programs that are implemented with Tensorflow and PyTorch, which are two commonly used AD frameworks.

(0 min) It is a well-known fact that nonconvex optimization is computationally intractable in the worst case. As a result, theoretical analysis of optimization algorithms such as gradient descent often focuses on local convergence to stationary points where the gradient norm is zero or negligible. In this work, we examine the disconnect between the existing theoretical analysis of gradient-based algorithms and actual practice. Specifically, we provide numerical evidence that in large-scale neural network training, such as in ImageNet, ResNet, and WT103 + TransformerXL models, the Neural Network weight variables do not converge to stationary points where the gradient of the loss function vanishes. Remarkably, however, we observe that while weights do not converge to stationary points, the value of the loss function converges. Inspired by this observation, we propose a new perspective based on ergodic theory of dynamical systems. We prove convergence of the distribution of weight values to an approximate invariant measure (without smoothness assumptions) that explains this phenomenon. We further discuss how this perspective can better align the theory with empirical observations.

(0 min) Modern neural network architectures can leverage large amounts of data to generalize well within the training distribution. However, they are less capable of systematic generalization to data drawn from unseen but related distributions, a feat that is hypothesized to require compositional reasoning and reuse of knowledge. In this work, we present Neural Interpreters, an architecture that factorizes inference in a self-attention network as a system of modules, which we call \emph{functions}. Inputs to the model are routed through a sequence of functions in a way that is end-to-end learned. The proposed architecture can flexibly compose computation along width and depth, and lends itself well to capacity extension after training. To demonstrate the versatility of Neural Interpreters, we evaluate it in two distinct settings: image classification and visual abstract reasoning on Raven Progressive Matrices. In the former, we show that Neural Interpreters perform on par with the vision transformer using fewer parameters, while being transferrable to a new task in a sample efficient manner. In the latter, we find that Neural Interpreters are competitive with respect to the state-of-the-art in terms of systematic generalization

(0 min) Stabilizing an unknown control system is one of the most fundamental problems in control systems engineering. In this paper, we provide a simple, model-free algorithm for stabilizing fully observed dynamical systems. While model-free methods have become increasingly popular in practice due to their simplicity and flexibility, stabilization via direct policy search has received surprisingly little attention. Our algorithm proceeds by solving a series of discounted LQR problems, where the discount factor is gradually increased. We prove that this method efficiently recovers a stabilizing controller for linear systems, and for smooth, nonlinear systems within a neighborhood of their equilibria. Our approach overcomes a significant limitation of prior work, namely the need for a pre-given stabilizing control policy. We empirically evaluate the effectiveness of our approach on common control benchmarks.

(0 min) Pre-trained language models (PLMs) aim to learn universal language representations by conducting self-supervised training tasks on large-scale corpora. Since PLMs capture word semantics in different contexts, the quality of word representations highly depends on word frequency, which usually follows a heavy-tailed distributions in the pre-training corpus. Therefore, the embeddings of rare words on the tail are usually poorly optimized. In this work, we focus on enhancing language model pre-training by leveraging definitions of the rare words in dictionaries (e.g., Wiktionary). To incorporate a rare word definition as a part of input, we fetch its definition from the dictionary and append it to the end of the input text sequence. In addition to training with the masked language modeling objective, we propose two novel self-supervised pre-training tasks on word and sentence-level alignment between input text sequence and rare word definitions to enhance language modeling representation with dictionary. We evaluate the proposed Dict-BERT model on the language understanding benchmark GLUE and eight specialized domain benchmark datasets. Extensive experiments demonstrate that Dict-BERT can significantly improve the understanding of rare words and boost model performance on various NLP downstream tasks.

(0 min) Weakly supervised semantic segmentation produces pixel-level localization from class labels; however, a classifier trained on such labels is likely to focus on a small discriminative region of the target object. We interpret this phenomenon using the information bottleneck principle: the final layer of a deep neural network, activated by the sigmoid or softmax activation functions, causes an information bottleneck, and as a result, only a subset of the task-relevant information is passed on to the output. We first support this argument through a simulated toy experiment and then propose a method to reduce the information bottleneck by removing the last activation function. In addition, we introduce a new pooling method that further encourages the transmission of information from non-discriminative regions to the classification. Our experimental evaluations demonstrate that this simple modification significantly improves the quality of localization maps on both the PASCAL VOC 2012 and MS COCO 2014 datasets, exhibiting a new state-of-the-art performance for weakly supervised semantic segmentation. The code is available at: https://github.com/jbeomlee93/RIB.

(0 min) We give simpler, sparser, and faster algorithms for differentially private fine-tuning of large-scale pre-trained language models, which achieve the state-of-the-art privacy versus utility tradeoffs on many standard NLP tasks. We propose a meta-framework for this problem, inspired by the recent success of highly parameter-efficient methods for fine-tuning. Our experiments show that differentially private adaptations of these approaches outperform previous private algorithms in three important dimensions: utility, privacy, and the computational and memory cost of private training. On many commonly studied datasets, the utility of private models approaches that of non-private models. For example, on the MNLI dataset we achieve an accuracy of $87.8\%$ using RoBERTa-Large and $83.5\%$ using RoBERTa-Base with a privacy budget of $\epsilon = 6.7$. In comparison, absent privacy constraints, RoBERTa-Large achieves an accuracy of $90.2\%$. Our findings are similar for natural language generation tasks. Privately fine-tuning with DART, GPT-2-Small, GPT-2-Medium, GPT-2-Large, and GPT-2-XL achieve BLEU scores of 38.5, 42.0, 43.1, and 43.8 respectively (privacy budget of $\epsilon = 6.8,\delta=$ 1e-5) whereas the non-private baseline is $48.1$. All our experiments suggest that larger models are better suited for private fine-tuning: while they are well known to achieve superior accuracy non-privately, we find that they also better maintain their accuracy when privacy is introduced.

(0 min) The Jobs shop Scheduling Problem (JSP) is a canonical combinatorial optimization problem that is routinely solved for a variety of industrial purposes. It models the optimal scheduling of multiple sequences of tasks, each under a fixed order of operations, in which individual tasks require exclusive access to a predetermined resource for a specified processing time. The problem is NP-hard and computationally challenging even for medium-sized instances. Motivated by the increased stochasticity in production chains, this paper explores a deep learning approach to deliver efficient and accurate approximations to the JSP. In particular, this paper proposes the design of a deep neural network architecture to exploit the problem structure, its integration with Lagrangian duality to capture the problem constraints, and a post-processing optimization to guarantee solution feasibility.The resulting method, called JSP-DNN, is evaluated on hard JSP instances from the JSPLIB benchmark library. Computational results show that JSP-DNN can produce JSP approximations of high quality at negligible computational costs.

(0 min) Molecular design and synthesis planning are two critical steps in the process of molecular discovery that we propose to formulate as a single shared task of conditional synthetic pathway generation. We report an amortized approach to generate synthetic pathways as a Markov decision process conditioned on a target molecular embedding. This approach allows us to conduct synthesis planning in a bottom-up manner and design synthesizable molecules by decoding from optimized conditional codes, demonstrating the potential to solve both problems of design and synthesis simultaneously. The approach leverages neural networks to probabilistically model the synthetic trees, one reaction step at a time, according to reactivity rules encoded in a discrete action space of reaction templates. We train these networks on hundreds of thousands of artificial pathways generated from a pool of purchasable compounds and a list of expert-curated templates. We validate our method with (a) the recovery of molecules using conditional generation, (b) the identification of synthesizable structural analogs, and (c) the optimization of molecular structures given oracle functions relevant to drug discovery.

(0 min) We study non-convex subgradient flows for training two-layer ReLU neural networks from a convex geometry and duality perspective. We characterize the implicit bias of unregularized non-convex gradient flow as convex regularization of an equivalent convex model. We then show that the limit points of non-convex subgradient flows can be identified via primal-dual correspondence in this convex optimization problem. Moreover, we derive a sufficient condition on the dual variables which ensures that the stationary points of the non-convex objective are the KKT points of the convex objective, thus proving convergence of non-convex gradient flows to the global optimum. For a class of regular training data distributions such as orthogonal separable data, we show that this sufficient condition holds. Therefore, non-convex gradient flows in fact converge to optimal solutions of a convex optimization problem. We present numerical results verifying the predictions of our theory for non-convex subgradient descent.

(0 min) The new coronavirus (known as COVID-19) was first identified in Wuhan and quickly spread worldwide, wreaking havoc on the economy and people's everyday lives. Fever, cough, sore throat, headache, exhaustion, muscular aches, and difficulty breathing are all typical symptoms of COVID-19. A reliable detection technique is needed to identify affected individuals and care for them in the early stages of COVID-19 and reduce the virus's transmission. The most accessible method for COVID-19 identification is RT-PCR; however, due to its time commitment and false-negative results, alternative options must be sought. Indeed, compared to RT-PCR, chest CT scans and chest X-ray images provide superior results. Because of the scarcity and high cost of CT scan equipment, X-ray images are preferable for screening. In this paper, a pre-trained network, DenseNet169, was employed to extract features from X-ray images. Features were chosen by a feature selection method (ANOVA) to reduce computations and time complexity while overcoming the curse of dimensionality to improve predictive accuracy. Finally, selected features were classified by XGBoost. The ChestX-ray8 dataset, which was employed to train and evaluate the proposed method. This method reached 98.72% accuracy for two-class classification (COVID-19, healthy) and 92% accuracy for three-class classification (COVID-19, healthy, pneumonia).

(0 min) We provide explicit bounds on the number of sample points required to estimate tangent spaces and intrinsic dimensions of (smooth, compact) Euclidean submanifolds via local principal component analysis. Our approach directly estimates covariance matrices locally, which simultaneously allows estimating both the tangent spaces and the intrinsic dimension of a manifold. The key arguments involve a matrix concentration inequality, a Wasserstein bound for flattening a manifold, and a Lipschitz relation for the covariance matrix with respect to the Wasserstein distance.

(0 min) With the explosive increase of big data, training a Machine Learning (ML) model becomes a computation-intensive workload, which would take days or even weeks. Thus, model reuse has received attention in the ML community, where it is called transfer learning. Transfer learning avoids training a new model from scratch by transferring knowledge from a source task to a target task. Existing transfer learning methods mostly focus on how to improve the performance of the target task through a specific source model, but assume that the source model is given. As many source models are available, it is difficult for data scientists to select the best source model for the target task manually. Hence, how to efficiently select a suitable source model for model reuse is still an unsolved problem. In this paper, we propose SMS, an effective, efficient and flexible source model selection framework. SMS is effective even when source and target datasets have significantly different data labels, is flexible to support source models with any type of structure, and is efficient to avoid any training process. For each source model, SMS first vectorizes the samples in the target dataset into soft labels by directly applying this model to the target dataset, then uses Gaussian distributions to fit the clusters of soft labels, and finally measures its distinguishing ability using Gaussian mixture-based metric. Moreover, we present an improved SMS (I-SMS), which decreases the output number of source model. I-SMS can significantly reduce the selection time while retaining the selection performance of SMS. Extensive experiments on a range of practical model reuse workloads demonstrate the effectiveness and efficiency of SMS.

(2 min) Conventional vocoders are commonly used as analysis tools to provide interpretable features for downstream tasks such as speech synthesis and voice conversion. They are built under certain assumptions about the signals following signal processing principle, therefore, not easily generalizable to different audio, for example, from speech to singing. In this paper, we propose a deep neural analyzer, denoted as DeepA - a neural vocoder that extracts F0 and timbre/aperiodicity encoding from the input speech that emulate those defined in conventional vocoders. Therefore, the resulting parameters are more interpretable than other latent neural representations. At the same time, as the deep neural analyzer is learnable, it is expected to be more accurate for signal reconstruction and manipulation, and generalizable from speech to singing. The proposed neural analyzer is built based on a variational autoencoder (VAE) architecture. We show that DeepA improves F0 estimation over the conventional vocoder (WORLD). To our best knowledge, this is the first study dedicated to the development of a neural framework for extracting learnable vocoder-like parameters.

(2 min) Latent-space interpolation is commonly used to demonstrate the generalization ability of deep latent variable models. Various algorithms have been proposed to calculate the best trajectory between two encodings in the latent space. In this work, we show how data labeled with semantically continuous attributes can be utilized to conduct a quantitative evaluation of latent-space interpolation algorithms, for variational autoencoders. Our framework can be used to complement the standard qualitative comparison, and also enables evaluation for domains (such as graph) in which the visualization is difficult. Interestingly, our experiments reveal that the superiority of interpolation algorithms could be domain-dependent. While normalised interpolation works best for the image domain, spherical linear interpolation achieves the best performance in the graph domain. Next, we propose a simple-yet-effective method to restrict the latent space via a bottleneck structure in the encoder. We find that all interpolation algorithms evaluated in this work can benefit from this restriction. Finally, we conduct interpolation-aware training with the labeled attributes, and show that this explicit supervision can improve the interpolation performance.

(0 min) The electric grid is a key enabling infrastructure for the ambitious transition towards carbon neutrality as we grapple with climate change. With deepening penetration of renewable energy resources and electrified transportation, the reliable and secure operation of the electric grid becomes increasingly challenging. In this paper, we present PSML, a first-of-its-kind open-access multi-scale time-series dataset, to aid in the development of data-driven machine learning (ML) based approaches towards reliable operation of future electric grids. The dataset is generated through a novel transmission + distribution (T+D) co-simulation designed to capture the increasingly important interactions and uncertainties of the grid dynamics, containing electric load, renewable generation, weather, voltage and current measurements at multiple spatio-temporal scales. Using PSML, we provide state-of-the-art ML baselines on three challenging use cases of critical importance to achieve: (i) early detection, accurate classification and localization of dynamic disturbance events; (ii) robust hierarchical forecasting of load and renewable energy with the presence of uncertainties and extreme events; and (iii) realistic synthetic generation of physical-law-constrained measurement time series. We envision that this dataset will enable advances for ML in dynamic systems, while simultaneously allowing ML researchers to contribute towards carbon-neutral electricity and mobility.

(2 min) Federated learning, in which training data is distributed among users and never shared, has emerged as a popular approach to privacy-preserving machine learning. Cryptographic techniques such as secure aggregation are used to aggregate contributions, like a model update, from all users. A robust technique for making such aggregates differentially private is to exploit infinite divisibility of the Laplace distribution, namely, that a Laplace distribution can be expressed as a sum of i.i.d. noise shares from a Gamma distribution, one share added by each user. However, Laplace noise is known to have suboptimal error in the low privacy regime for $\varepsilon$-differential privacy, where $\varepsilon > 1$ is a large constant. In this paper we present the first infinitely divisible noise distribution for real-valued data that achieves $\varepsilon$-differential privacy and has expected error that decreases exponentially with $\varepsilon$.

(2 min) Graph neural networks (GNNs) have been used extensively for addressing problems in drug design and discovery. Both ligand and target molecules are represented as graphs with node and edge features encoding information about atomic elements and bonds respectively. Although existing deep learning models perform remarkably well at predicting physicochemical properties and binding affinities, the generation of new molecules with optimized properties remains challenging. Inherently, most GNNs perform poorly in whole-graph representation due to the limitations of the message-passing paradigm. Furthermore, step-by-step graph generation frameworks that use reinforcement learning or other sequential processing can be slow and result in a high proportion of invalid molecules with substantial post-processing needed in order to satisfy the principles of stoichiometry. To address these issues, we propose a representation-first approach to molecular graph generation. We guide the latent representation of an autoencoder by capturing graph structure information with the geometric scattering transform and apply penalties that structure the representation also by molecular properties. We show that this highly structured latent space can be directly used for molecular graph generation by the use of a GAN. We demonstrate that our architecture learns meaningful representations of drug datasets and provides a platform for goal-directed drug synthesis.

(2 min) Pre-trained language models in the past years have shown exponential growth in model parameters and compute time. ELECTRA is a novel approach for improving the compute efficiency of pre-trained language models (e.g. BERT) based on masked language modeling (MLM) by addressing the sample inefficiency problem with the replaced token detection (RTD) task. Our work proposes adaptive early exit strategy to maximize the efficiency of the pre-training process by relieving the model's subsequent layers of the need to process latent features by leveraging earlier layer representations. Moreover, we evaluate an initial approach to the problem that has not succeeded in maintaining the accuracy of the model while showing a promising compute efficiency by thoroughly investigating the necessity of the generator module of ELECTRA.

(2 min) A central task of a Disc Jockey (DJ) is to create a mixset of mu-sic with seamless transitions between adjacent tracks. In this paper, we explore a data-driven approach that uses a generative adversarial network to create the song transition by learning from real-world DJ mixes. In particular, the generator of the model uses two differentiable digital signal processing components, an equalizer (EQ) and a fader, to mix two tracks selected by a data generation pipeline. The generator has to set the parameters of the EQs and fader in such away that the resulting mix resembles real mixes created by humanDJ, as judged by the discriminator counterpart. Result of a listening test shows that the model can achieve competitive results compared with a number of baselines.

(2 min) This paper aims to automatically detect COVID-19 patients by analysing the acoustic information embedded in coughs. COVID-19 affects the respiratory system, and, consequently, respiratory-related signals have the potential to contain salient information for the task at hand. We focus on analysing the spectrogram representations of coughing samples with the aim to investigate whether COVID-19 alters the frequency content of these signals. Furthermore, this work also assesses the impact of gender in the automatic detection of COVID-19. To extract deep learnt representations of the spectrograms, we compare the performance of a cough-specific, and a Resnet18 pre-trained Convolutional Neural Network (CNN). Additionally, our approach explores the use of contextual attention, so the model can learn to highlight the most relevant deep learnt features extracted by the CNN. We conduct our experiments on the dataset released for the Cough Sound Track of the DiCOVA 2021 Challenge. The best performance on the test set is obtained using the Resnet18 pre-trained CNN with contextual attention, which scored an Area Under the Curve (AUC) of 70.91 at 80% sensitivity.

(2 min) Recommendation systems play an important role in today's digital world. They have found applications in various applications such as music platforms, e.g., Spotify, and movie streaming services, e.g., Netflix. Less research effort has been devoted to physical exercise recommendation systems. Sedentary lifestyles have become the major driver of several diseases as well as healthcare costs. In this paper, we develop a recommendation system for daily exercise activities to users based on their history, profile and similar users. The developed recommendation system uses a deep recurrent neural network with user-profile attention and temporal attention mechanisms. Moreover, exercise recommendation systems are significantly different from streaming recommendation systems in that we are not able to collect click feedback from the participants in exercise recommendation systems. Thus, we propose a real-time, expert-in-the-loop active learning procedure. The active learners calculate the uncertainty of the recommender at each time step for each user and ask an expert for a recommendation when the certainty is low. In this paper, we derive the probability distribution function of marginal distance, and use it to determine when to ask experts for feedback. Our experimental results on a mHealth dataset show improved accuracy after incorporating the real-time active learner with the recommendation system.

(2 min) Deep neural networks can be roughly divided into deterministic neural networks and stochastic neural networks.The former is usually trained to achieve a mapping from input space to output space via maximum likelihood estimation for the weights, which leads to deterministic predictions during testing. In this way, a specific weights set is estimated while ignoring any uncertainty that may occur in the proper weight space. The latter introduces randomness into the framework, either by assuming a prior distribution over model parameters (i.e. Bayesian Neural Networks) or including latent variables (i.e. generative models) to explore the contribution of latent variables for model predictions, leading to stochastic predictions during testing. Different from the former that achieves point estimation, the latter aims to estimate the prediction distribution, making it possible to estimate uncertainty, representing model ignorance about its predictions. We claim that conventional deterministic neural network based dense prediction tasks are prone to overfitting, leading to over-confident predictions, which is undesirable for decision making. In this paper, we investigate stochastic neural networks and uncertainty estimation techniques to achieve both accurate deterministic prediction and reliable uncertainty estimation. Specifically, we work on two types of uncertainty estimations solutions, namely ensemble based methods and generative model based methods, and explain their pros and cons while using them in fully/semi/weakly-supervised framework. Due to the close connection between uncertainty estimation and model calibration, we also introduce how uncertainty estimation can be used for deep model calibration to achieve well-calibrated models, namely dense model calibration. Code and data are available at https://github.com/JingZhang617/UncertaintyEstimation.

(2 min) Autonomous driving is of great interest in both research and industry. The high cost has been one of the major roadblocks that slow down the development and adoption of autonomous driving in practice. This paper, for the first-time, shows that it is possible to run level-4 (i.e., fully autonomous driving) software on a single off-the-shelf card (Jetson AGX Xavier) for less than $1k, an order of magnitude less than the state-of-the-art systems, while meeting all the requirements of latency. The success comes from the resolution of some important issues shared by existing practices through a series of measures and innovations. The study overturns the common perceptions of the computing resources required by level-4 autonomous driving, points out a promising path for the industry to lower the cost, and suggests a number of research opportunities for rethinking the architecture, software design, and optimizations of autonomous driving.

(2 min) Solving for the lowest energy eigenstate of the many-body Schrodinger equation is a cornerstone problem that hinders understanding of a variety of quantum phenomena. The difficulty arises from the exponential nature of the Hilbert space which casts the governing equations as an eigenvalue problem of exponentially large, structured matrices. Variational methods approach this problem by searching for the best approximation within a lower-dimensional variational manifold. In this work we use graph neural networks to define a structured variational manifold and optimize its parameters to find high quality approximations of the lowest energy solutions on a diverse set of Heisenberg Hamiltonians. Using graph networks we learn distributed representations that by construction respect underlying physical symmetries of the problem and generalize to problems of larger size. Our approach achieves state-of-the-art results on a set of quantum many-body benchmark problems and works well on problems whose solutions are not positive-definite. The discussed techniques hold promise of being a useful tool for studying quantum many-body systems and providing insights into optimization and implicit modeling of exponentially-sized objects.

(2 min) We propose a novel approach to translate unpaired contrast computed tomography (CT) scans to non-contrast CT scans and the other way around. Solving this task has two important applications: (i) to automatically generate contrast CT scans for patients for whom injecting contrast substance is not an option, and (ii) to enhance alignment between contrast and non-contrast CT by reducing the differences induced by the contrast substance before registration. Our approach is based on cycle-consistent generative adversarial convolutional transformers, for short, CyTran. Our neural model can be trained on unpaired images, due to the integration of a cycle-consistency loss. To deal with high-resolution images, we design a hybrid architecture based on convolutional and multi-head attention layers. In addition, we introduce a novel data set, Coltea-Lung-CT-100W, containing 3D triphasic lung CT scans (with a total of 37,290 images) collected from 100 female patients. Each scan contains three phases (non-contrast, early portal venous, and late arterial), allowing us to perform experiments to compare our novel approach with state-of-the-art methods for image style transfer. Our empirical results show that CyTran outperforms all competing methods. Moreover, we show that CyTran can be employed as a preliminary step to improve a state-of-the-art medical image alignment method. We release our novel model and data set as open source at: https://github.com/ristea/cycle-transformer.

(3 min) Dynamic Mode Decomposition (DMD) is an unsupervised machine learning method that has attracted considerable attention in recent years owing to its equation-free structure, ability to easily identify coherent spatio-temporal structures in data, and effectiveness in providing reasonably accurate predictions for certain problems. Despite these successes, the application of DMD to certain problems featuring highly nonlinear transient dynamics remains challenging. In such cases, DMD may not only fail to provide acceptable predictions but may indeed fail to recreate the data in which it was trained, restricting its application to diagnostic purposes. For many problems in the biological and physical sciences, the structure of the system obeys a compartmental framework, in which the transfer of mass within the system moves within states. In these cases, the behavior of the system may not be accurately recreated by applying DMD to a single quantity within the system, as proper knowledge of the system dynamics, even for a single compartment, requires that the behavior of other compartments is taken into account in the DMD process. In this work, we demonstrate, theoretically and numerically, that, when performing DMD on a fully coupled PDE system with compartmental structure, one may recover useful predictive behavior, even when DMD performs poorly when acting compartment-wise. We also establish that important physical quantities, as mass conservation, are maintained in the coupled-DMD extrapolation. The mathematical and numerical analysis suggests that DMD may be a powerful tool when applied to this common class of problems. In particular, we show interesting numerical applications to a continuous delayed-SIRD model for Covid-19, and to a problem from additive manufacturing considering a nonlinear temperature field and the resulting change of material phase from powder, liquid, and solid states.

(0 min) Monitoring progress on the United Nations Sustainable Development Goals (SDGs) is important for both academic and non-academic organizations. Existing approaches to monitoring SDGs have focused on specific data types, namely, publications listed in proprietary research databases. We present the text2sdg R package, a user-friendly, open-source package that detects SDGs in any kind of text data using several different query systems from any text source. The text2sdg package thereby facilitates the monitoring of SDGs for a wide array of text sources and provides a much-needed basis for validating and improving extant methods to detect SDGs from text.

(0 min) Trained on the large corpus, pre-trained language models (PLMs) can capture different levels of concepts in context and hence generate universal language representations. They can benefit multiple downstream natural language processing (NLP) tasks. Although PTMs have been widely used in most NLP applications, especially for high-resource languages such as English, it is under-represented in Lao NLP research. Previous work on Lao has been hampered by the lack of annotated datasets and the sparsity of language resources. In this work, we construct a text classification dataset to alleviate the resource-scare situation of the Lao language. We additionally present the first transformer-based PTMs for Lao with four versions: BERT-small, BERT-base, ELECTRA-small and ELECTRA-base, and evaluate it over two downstream tasks: part-of-speech tagging and text classification. Experiments demonstrate the effectiveness of our Lao models. We will release our models and datasets to the community, hoping to facilitate the future development of Lao NLP applications.

(0 min) Dialectometry is a discipline devoted to studying the variations of a language around a geographical region. One of their goals is the creation of linguistic atlases capturing the similarities and differences of the language under study around the area in question. For instance, Spanish is one of the most spoken languages across the world, but not necessarily Spanish is written and spoken in the same way in different countries. This manuscript presents a broad analysis describing lexical and semantic relationships among 26 Spanish-speaking countries around the globe. For this study, we analyze four-year of the Twitter geotagged public stream to provide an extensive survey of the Spanish language vocabularies of different countries, its distributions, semantic usage of terms, and emojis. We also offer open regional word-embedding resources for Spanish Twitter to help other researchers and practitioners take advantage of regionalized models.

(0 min) Data processing is an important step in various natural language processing tasks. As the commonly used datasets in named entity recognition contain only a limited number of samples, it is important to obtain additional labeled data in an efficient and reliable manner. A common practice is to utilize large monolingual unlabeled corpora. Another popular technique is to create synthetic data from the original labeled data (data augmentation). In this work, we investigate the impact of these two methods on the performance of three different named entity recognition tasks.

(0 min) When training and evaluating machine learning models on a large number of tasks, it is important to not only look at average task accuracy -- which may be biased by easy or redundant tasks -- but also worst-case accuracy (i.e. the performance on the task with the lowest accuracy). In this work, we show how to use techniques from the distributionally robust optimization (DRO) literature to improve worst-case performance in multitask learning. We highlight several failure cases of DRO when applied off-the-shelf and present an improved method, Lookahead-DRO (L-DRO), which mitigates these issues. The core idea of L-DRO is to anticipate the interaction between tasks during training in order to choose a dynamic re-weighting of the various task losses, which will (i) lead to minimal worst-case loss and (ii) train on as many tasks as possible. After demonstrating the efficacy of L-DRO on a small controlled synthetic setting, we evaluate it on two realistic benchmarks: a multitask version of the CIFAR-100 image classification dataset and a large-scale multilingual language modeling experiment. Our empirical results show that L-DRO achieves a better trade-off between average and worst-case accuracy with little computational overhead compared to several strong baselines.

(0 min) We present work on summarising deliberative processes for non-English languages. Unlike commonly studied datasets, such as news articles, this deliberation dataset reflects difficulties of combining multiple narratives, mostly of poor grammatical quality, in a single text. We report an extensive evaluation of a wide range of abstractive summarisation models in combination with an off-the-shelf machine translation model. Texts are translated into English, summarised, and translated back to the original language. We obtain promising results regarding the fluency, consistency and relevance of the summaries produced. Our approach is easy to implement for many languages for production purposes by simply changing the translation model.

(0 min) Multi-modal reasoning systems rely on a pre-trained object detector to extract regions of interest from the image. However, this crucial module is typically used as a black box, trained independently of the downstream task and on a fixed vocabulary of objects and attributes. This makes it challenging for such systems to capture the long tail of visual concepts expressed in free form text. In this paper we propose MDETR, an end-to-end modulated detector that detects objects in an image conditioned on a raw text query, like a caption or a question. We use a transformer-based architecture to reason jointly over text and image by fusing the two modalities at an early stage of the model. We pre-train the network on 1.3M text-image pairs, mined from pre-existing multi-modal datasets having explicit alignment between phrases in text and objects in the image. We then fine-tune on several downstream tasks such as phrase grounding, referring expression comprehension and segmentation, achieving state-of-the-art results on popular benchmarks. We also investigate the utility of our model as an object detector on a given label set when fine-tuned in a few-shot setting. We show that our pre-training approach provides a way to handle the long tail of object categories which have very few labelled instances. Our approach can be easily extended for visual question answering, achieving competitive performance on GQA and CLEVR. The code and models are available at https://github.com/ashkamath/mdetr.

(0 min) A major challenge in structured prediction is to represent the interdependencies within output structures. When outputs are structured as sequences, linear-chain conditional random fields (CRFs) are a widely used model class which can learn \textit{local} dependencies in the output. However, the CRF's Markov assumption makes it impossible for CRFs to represent distributions with \textit{nonlocal} dependencies, and standard CRFs are unable to respect nonlocal constraints of the data (such as global arity constraints on output labels). We present a generalization of CRFs that can enforce a broad class of constraints, including nonlocal ones, by specifying the space of possible output structures as a regular language $\mathcal{L}$. The resulting regular-constrained CRF (RegCCRF) has the same formal properties as a standard CRF, but assigns zero probability to all label sequences not in $\mathcal{L}$. Notably, RegCCRFs can incorporate their constraints during training, while related models only enforce constraints during decoding. We prove that constrained training is never worse than constrained decoding, and show empirically that it can be substantially better in practice. Additionally, we demonstrate a practical benefit on downstream tasks by incorporating a RegCCRF into a deep neural model for semantic role labeling, exceeding state-of-the-art results on a standard dataset.

(0 min) Neural Machine Translation (NMT) models are known to suffer from noisy inputs. To make models robust, we generate adversarial augmentation samples that attack the model and preserve the source-side semantic meaning at the same time. To generate such samples, we propose a doubly-trained architecture that pairs two NMT models of opposite translation directions with a joint loss function, which combines the target-side attack and the source-side semantic similarity constraint. The results from our experiments across three different language pairs and two evaluation metrics show that these adversarial samples improve the model robustness.

(0 min) Most of the deep learning-based speech enhancement models are learned in a supervised manner, which implies that pairs of noisy and clean speech are required during training. Consequently, several noisy speeches recorded in daily life cannot be used to train the model. Although certain unsupervised learning frameworks have also been proposed to solve the pair constraint, they still require clean speech or noise for training. Therefore, in this paper, we propose MetricGAN-U, which stands for MetricGAN-unsupervised, to further release the constraint from conventional unsupervised learning. In MetricGAN-U, only noisy speech is required to train the model by optimizing non-intrusive speech quality metrics. The experimental results verified that MetricGAN-U outperforms baselines in both objective and subjective metrics.

(2 min) Sports game summarization aims to generate news articles from live text commentaries. A recent state-of-the-art work, SportsSum, not only constructs a large benchmark dataset, but also proposes a two-step framework. Despite its great contributions, the work has three main drawbacks: 1) the noise existed in SportsSum dataset degrades the summarization performance; 2) the neglect of lexical overlap between news and commentaries results in low-quality pseudo-labeling algorithm; 3) the usage of directly concatenating rewritten sentences to form news limits its practicability. In this paper, we publish a new benchmark dataset SportsSum2.0, together with a modified summarization framework. In particular, to obtain a clean dataset, we employ crowd workers to manually clean the original dataset. Moreover, the degree of lexical overlap is incorporated into the generation of pseudo labels. Further, we introduce a reranker-enhanced summarizer to take into account the fluency and expressiveness of the summarized news. Extensive experiments show that our model outperforms the state-of-the-art baseline.

(2 min) Automatic Speech Recognition (ASR) is gaining special interest in Air Traffic Control (ATC). ASR allows transcribing the communications between air traffic controllers (ATCOs) and pilots. These transcriptions are used to extract ATC command types and named entities such as aircraft callsigns. One common problem is when the Speech Activity Detection (SAD) or diarization system fails and then two or more single speaker segments are in the same recording, jeopardizing the overall system's performance. We developed a system that combines the segmentation of a SAD module with a BERT-based model that performs Speaker Change Detection (SCD) and Speaker Role Identification (SRI) based on ASR transcripts (i.e., diarization + SRI). This research demonstrates on a real-life ATC test set that performing diarization directly on textual data surpass acoustic level diarization. The proposed model reaches up to ~0.90/~0.95 F1-score on ATCO/pilot for SRI on several test sets. The text-based diarization system brings a 27% relative improvement on Diarization Error Rate (DER) compared to standard acoustic-based diarization. These results were on ASR transcripts of a challenging ATC test set with an estimated ~13% word error rate, validating the approach's robustness even on noisy ASR transcripts.

(0 min) Majority voting and averaging are common approaches employed to resolve annotator disagreements and derive single ground truth labels from multiple annotations. However, annotators may systematically disagree with one another, often reflecting their individual biases and values, especially in the case of subjective tasks such as detecting affect, aggression, and hate speech. Annotator disagreements may capture important nuances in such tasks that are often ignored while aggregating annotations to a single ground truth. In order to address this, we investigate the efficacy of multi-annotator models. In particular, our multi-task based approach treats predicting each annotators' judgements as separate subtasks, while sharing a common learned representation of the task. We show that this approach yields same or better performance than aggregating labels in the data prior to training across seven different binary classification tasks. Our approach also provides a way to estimate uncertainty in predictions, which we demonstrate better correlate with annotation disagreements than traditional methods. Being able to model uncertainty is especially useful in deployment scenarios where knowing when not to make a prediction is important.

(0 min) Despite the recent advances in applying pre-trained language models to generate high-quality texts, generating long passages that maintain long-range coherence is yet challenging for these models. In this paper, we propose DiscoDVT, a discourse-aware discrete variational Transformer to tackle the incoherence issue. DiscoDVT learns a discrete variable sequence that summarizes the global structure of the text and then applies it to guide the generation process at each decoding step. To further embed discourse-aware information into the discrete latent representations, we introduce an auxiliary objective to model the discourse relations within the text. We conduct extensive experiments on two open story generation datasets and demonstrate that the latent codes learn meaningful correspondence to the discourse structures that guide the model to generate long texts with better long-range coherence.

(2 min) In 2020, COVID-19 became the chief concern of the world and is still reflected widely in all social networks. Each day, users post millions of tweets and comments on this subject, which contain significant implicit information about the public opinion. In this regard, a dataset of COVID-related tweets in English language is collected, which consists of more than two million tweets from March 23 to June 23 of 2020 to extract the feelings of the people in various countries in the early stages of this outbreak. To this end, first, we use a lexicon-based approach in conjunction with the GeoNames geographic database to label the tweets with their locations. Next, a method based on the recently introduced and widely cited RoBERTa model is proposed to analyze their sentimental content. After that, the trend graphs of the frequency of tweets as well as sentiments are produced for the world and the nations that were more engaged with COVID-19. Graph analysis shows that the frequency graphs of the tweets for the majority of nations are significantly correlated with the official statistics of the daily afflicted in them. Moreover, several implicit knowledge is extracted and discussed.

(2 min) This work demonstrates that using the objective with independence assumption for modelling the span probability $P(a_s,a_e) = P(a_s)P(a_e)$ of span starting at position $a_s$ and ending at position $a_e$ has adverse effects. Therefore we propose multiple approaches to modelling joint probability $P(a_s,a_e)$ directly. Among those, we propose a compound objective, composed from the joint probability while still keeping the objective with independence assumption as an auxiliary objective. We find that the compound objective is consistently superior or equal to other assumptions in exact match. Additionally, we identified common errors caused by the assumption of independence and manually checked the counterpart predictions, demonstrating the impact of the compound objective on the real examples. Our findings are supported via experiments with three extractive QA models (BIDAF, BERT, ALBERT) over six datasets and our code, individual results and manual analysis are available online.

(2 min) A common practice in building NLP datasets, especially using crowd-sourced annotations, involves obtaining multiple annotator judgements on the same data instances, which are then flattened to produce a single "ground truth" label or score, through majority voting, averaging, or adjudication. While these approaches may be appropriate in certain annotation tasks, such aggregations overlook the socially constructed nature of human perceptions that annotations for relatively more subjective tasks are meant to capture. In particular, systematic disagreements between annotators owing to their socio-cultural backgrounds and/or lived experiences are often obfuscated through such aggregations. In this paper, we empirically demonstrate that label aggregation may introduce representational biases of individual and group perspectives. Based on this finding, we propose a set of recommendations for increased utility and transparency of datasets for downstream use cases.

(2 min) This work presents a supervised method for generating a classifier model of the stances held by Chinese-speaking politicians and other Twitter users. Many previous works of political tweets prediction exist on English tweets, but to the best of our knowledge, this is the first work that builds prediction model on Chinese political tweets. It firstly collects data by scraping tweets of famous political figure and their related users. It secondly defines the political spectrum in two groups: the group that shows approvals to the Chinese Communist Party and the group that does not. Since there are not space between words in Chinese to identify the independent words, it then completes segmentation and vectorization by Jieba, a Chinese segmentation tool. Finally, it trains the data collected from political tweets and produce a classification model with high accuracy for understanding users' political stances from their tweets on Twitter.

(2 min) Video-text retrieval has many real-world applications such as media analytics, surveillance, and robotics. This paper presents the 1st place solution to the video retrieval track of the ICCV VALUE Challenge 2021. We present a simple yet effective approach to jointly tackle two video-text retrieval tasks (video retrieval and video corpus moment retrieval) by leveraging the model trained only on the video retrieval task. In addition, we create an ensemble model that achieves the new state-of-the-art performance on all four datasets (TVr, How2r, YouCook2r, and VATEXr) presented in the VALUE Challenge.

(2 min) In this paper, we study training of automatic speech recognition system in a weakly supervised setting where the order of words in transcript labels of the audio training data is not known. We train a word-level acoustic model which aggregates the distribution of all output frames using LogSumExp operation and uses a cross-entropy loss to match with the ground-truth words distribution. Using the pseudo-labels generated from this model on the training set, we then train a letter-based acoustic model using Connectionist Temporal Classification loss. Our system achieves 2.4%/5.3% on test-clean/test-other subsets of LibriSpeech, which is competitive with the supervised baseline's performance.

(2 min) Parsing spoken dialogue poses unique difficulties, including disfluencies and unmarked boundaries between sentence-like units. Previous work has shown that prosody can help with parsing disfluent speech (Tran et al. 2018), but has assumed that the input to the parser is already segmented into sentence-like units (SUs), which isn't true in existing speech applications. We investigate how prosody affects a parser that receives an entire dialogue turn as input (a turn-based model), instead of gold standard pre-segmented SUs (an SU-based model). In experiments on the English Switchboard corpus, we find that when using transcripts alone, the turn-based model has trouble segmenting SUs, leading to worse parse performance than the SU-based model. However, prosody can effectively replace gold standard SU boundaries: with prosody, the turn-based model performs as well as the SU-based model (90.79 vs. 90.65 F1 score, respectively), despite performing two tasks (SU segmentation and parsing) rather than one (parsing alone). Analysis shows that pitch and intensity features are the most important for this corpus, since they allow the model to correctly distinguish an SU boundary from a speech disfluency -- a distinction that the model otherwise struggles to make.

(2 min) Dialog is a core building block of human natural language interactions. It contains multi-party utterances used to convey information from one party to another in a dynamic and evolving manner. The ability to compare dialogs is beneficial in many real world use cases, such as conversation analytics for contact center calls and virtual agent design. We propose a novel adaptation of the edit distance metric to the scenario of dialog similarity. Our approach takes into account various conversation aspects such as utterance semantics, conversation flow, and the participants. We evaluate this new approach and compare it to existing document similarity measures on two publicly available datasets. The results demonstrate that our method outperforms the other approaches in capturing dialog flow, and is better aligned with the human perception of conversation similarity.

(2 min) ALFRED is a recently proposed benchmark that requires a model to complete tasks in simulated house environments specified by instructions in natural language. We hypothesize that key to success is accurately aligning the text modality with visual inputs. Motivated by this, we inspect how well existing models can align these modalities using our proposed intrinsic metric, boundary adherence score (BAS). The results show the previous models are indeed failing to perform proper alignment. To address this issue, we introduce approaches aimed at improving model alignment and demonstrate how improved alignment, improves end task performance.

(2 min) Temporal language grounding in videos aims to localize the temporal span relevant to the given query sentence. Previous methods treat it either as a boundary regression task or a span extraction task. This paper will formulate temporal language grounding into video reading comprehension and propose a Relation-aware Network (RaNet) to address it. This framework aims to select a video moment choice from the predefined answer set with the aid of coarse-and-fine choice-query interaction and choice-choice relation construction. A choice-query interactor is proposed to match the visual and textual information simultaneously in sentence-moment and token-moment levels, leading to a coarse-and-fine cross-modal interaction. Moreover, a novel multi-choice relation constructor is introduced by leveraging graph convolution to capture the dependencies among video moment choices for the best choice selection. Extensive experiments on ActivityNet-Captions, TACoS, and Charades-STA demonstrate the effectiveness of our solution. Codes will be released soon.

(2 min) Public transport agencies use social media as an essential tool for communicating mobility incidents to passengers. However, while the short term, day-to-day information about transport phenomena is usually posted in social media with low latency, its availability is short term as the content is rarely made an aggregated form. Social media communication of transport phenomena usually lacks GIS annotations as most social media platforms do not allow attaching non-POI GPS coordinates to posts. As a result, the analysis of transport phenomena information is minimal. We collected three years of social media posts of a polish public transport company with user comments. Through exploration, we infer a six-class transport information typology. We successfully build an information type classifier for social media posts, detect stop names in posts, and relate them to GPS coordinates, obtaining a spatial understanding of long-term aggregated phenomena. We show that our approach enables citizen science and use it to analyze the impact of three years of infrastructure incidents on passenger mobility, and the sentiment and reaction scale towards each of the events. All these results are achieved for Polish, an under-resourced language when it comes to spatial language understanding, especially in social media contexts. To improve the situation, we released two of our annotated data sets: social media posts with incident type labels and matched stop names and social media comments with the annotated sentiment. We also opensource the experimental codebase.

(2 min) We adopt an evolutionary view on language change in which cognitive factors (in addition to social ones) affect the fitness of words and their success in the linguistic ecosystem. Specifically, we propose a variety of psycholinguistic factors -- semantic, distributional, and phonological -- that we hypothesize are predictive of lexical decline, in which words greatly decrease in frequency over time. Using historical data across three languages (English, French, and German), we find that most of our proposed factors show a significant difference in the expected direction between each curated set of declining words and their matched stable words. Moreover, logistic regression analyses show that semantic and distributional factors are significant in predicting declining words. Further diachronic analysis reveals that declining words tend to decrease in the diversity of their lexical contexts over time, gradually narrowing their 'ecological niches'.

(2 min) The Transformer architecture has been well adopted as a dominant architecture in most sequence transduction tasks including automatic speech recognition (ASR), since its attention mechanism excels in capturing long-range dependencies. While models built solely upon attention can be better parallelized than regular RNN, a novel network architecture, SRU++, was recently proposed. By combining the fast recurrence and attention mechanism, SRU++ exhibits strong capability in sequence modeling and achieves near-state-of-the-art results in various language modeling and machine translation tasks with improved compute efficiency. In this work, we present the advantages of applying SRU++ in ASR tasks by comparing with Conformer across multiple ASR benchmarks and study how the benefits can be generalized to long-form speech inputs. On the popular LibriSpeech benchmark, our SRU++ model achieves 2.0% / 4.7% WER on test-clean / test-other, showing competitive performances compared with the state-of-the-art Conformer encoder under the same set-up. Specifically, SRU++ can surpass Conformer on long-form speech input with a large margin, based on our analysis.

(2 min) Training neural text-to-speech (TTS) models for a new speaker typically requires several hours of high quality speech data. Prior works on voice cloning attempt to address this challenge by adapting pre-trained multi-speaker TTS models for a new voice, using a few minutes of speech data of the new speaker. However, publicly available large multi-speaker datasets are often noisy, thereby resulting in TTS models that are not suitable for use in products. We address this challenge by proposing transfer-learning guidelines for adapting high quality single-speaker TTS models for a new speaker, using only a few minutes of speech data. We conduct an extensive study using different amounts of data for a new speaker and evaluate the synthesized speech in terms of naturalness and voice/style similarity to the target speaker. We find that fine-tuning a single-speaker TTS model on just 30 minutes of data, can yield comparable performance to a model trained from scratch on more than 27 hours of data for both male and female target speakers.

(2 min) Pre-trained language models (PLMs) have been the de facto paradigm for most natural language processing (NLP) tasks. This also benefits biomedical domain: researchers from informatics, medicine, and computer science (CS) communities propose various PLMs trained on biomedical datasets, e.g., biomedical text, electronic health records, protein, and DNA sequences for various biomedical tasks. However, the cross-discipline characteristics of biomedical PLMs hinder their spreading among communities; some existing works are isolated from each other without comprehensive comparison and discussions. It expects a survey that not only systematically reviews recent advances of biomedical PLMs and their applications but also standardizes terminology and benchmarks. In this paper, we summarize the recent progress of pre-trained language models in the biomedical domain and their applications in biomedical downstream tasks. Particularly, we discuss the motivations and propose a taxonomy of existing biomedical PLMs. Their applications in biomedical downstream tasks are exhaustively discussed. At last, we illustrate various limitations and future trends, which we hope can provide inspiration for the future research of the research community.

(2 min) In knowledge graph embedding, the theoretical relationship between the softmax cross-entropy and negative sampling loss functions has not been investigated. This makes it difficult to fairly compare the results of the two different loss functions. We attempted to solve this problem by using the Bregman divergence to provide a unified interpretation of the softmax cross-entropy and negative sampling loss functions. Under this interpretation, we can derive theoretical findings for fair comparison. Experimental results on the FB15k-237 and WN18RR datasets show that the theoretical findings are valid in practical settings.

(2 min) Training machines to understand natural language and interact with humans is an elusive and essential task of artificial intelligence. A diversity of dialogue systems has been designed with the rapid development of deep learning techniques, especially the recent pre-trained language models (PrLMs). Among these studies, the fundamental yet challenging type of task is dialogue comprehension whose role is to teach the machines to read and comprehend the dialogue context before responding. In this paper, we review the previous methods from the technical perspective of dialogue modeling for the dialogue comprehension task. We summarize the characteristics and challenges of dialogue comprehension in contrast to plain-text reading comprehension. Then, we discuss three typical patterns of dialogue modeling. In addition, we categorize dialogue-related pre-training techniques which are employed to enhance PrLMs in dialogue scenarios. Finally, we highlight the technical advances in recent years and point out the lessons from the empirical analysis and the prospects towards a new frontier of researches.

(2 min) Transformer-based models have proven to be powerful in many natural language, computer vision, and speech recognition applications. It is expensive to train these types of models due to unfixed input length, complex computation, and large numbers of parameters. Existing systems either only focus on efficient inference or optimize only BERT-like encoder models. In this paper, we present LightSeq, a system for efficient training of Transformer-based models on GPUs. We propose a series of GPU optimization techniques tailored to computation flow and memory access patterns of neural layers in Transformers. LightSeq supports a variety of network architectures, including BERT (encoder-only), GPT (decoder-only), and Transformer (encoder-decoder). Our experiments on GPUs with varying models and datasets show that LightSeq is 1.4-3.5x faster than previous systems. In particular, it gains 308% training speedup compared with existing systems on a large public machine translation benchmark (WMT14 English-German).

(2 min) The task of learning from only a few examples (called a few-shot setting) is of key importance and relevance to a real-world setting. For question answering (QA), the current state-of-the-art pre-trained models typically need fine-tuning on tens of thousands of examples to obtain good results. Their performance degrades significantly in a few-shot setting (< 100 examples). To address this, we propose a simple fine-tuning framework that leverages pre-trained text-to-text models and is directly aligned with their pre-training framework. Specifically, we construct the input as a concatenation of the question, a mask token representing the answer span and a context. Given this input, the model is fine-tuned using the same objective as that of its pre-training objective. Through experimental studies on various few-shot configurations, we show that this formulation leads to significant gains on multiple QA benchmarks (an absolute gain of 34.2 F1 points on average when there are only 16 training examples). The gains extend further when used with larger models (Eg:- 72.3 F1 on SQuAD using BART-large with only 32 examples) and translate well to a multilingual setting . On the multilingual TydiQA benchmark, our model outperforms the XLM-Roberta-large by an absolute margin of upto 40 F1 points and an average of 33 F1 points in a few-shot setting (<= 64 training examples). We conduct detailed ablation studies to analyze factors contributing to these gains.

(2 min) Inspired by the notion of Center of Mass in physics, an extension called Semantic Center of Mass (SCOM) is proposed, and used to discover the abstract "topic" of a document. The notion is under a framework model called Understanding Map Supervised Topic Model (UM-S-TM). The devise aim of UM-S-TM is to let both the document content and a semantic network -- specifically, Understanding Map -- play a role, in interpreting the meaning of a document. Based on different justifications, three possible methods are devised to discover the SCOM of a document. Some experiments on artificial documents and Understanding Maps are conducted to test their outcomes. In addition, its ability of vectorization of documents and capturing sequential information are tested. We also compared UM-S-TM with probabilistic topic models like Latent Dirichlet Allocation (LDA) and probabilistic Latent Semantic Analysis (pLSA).

(2 min) Neural models for the various flavours of morphological inflection tasks have proven to be extremely accurate given ample labeled data -- data that may be slow and costly to obtain. In this work we aim to overcome this annotation bottleneck by bootstrapping labeled data from a seed as little as {\em five} labeled paradigms, accompanied by a large bulk of unlabeled text. Our approach exploits different kinds of regularities in morphological systems in a two-phased setup, where word tagging based on {\em analogies} is followed by word pairing based on {\em distances}. We experiment with the Paradigm Cell Filling Problem over eight typologically different languages, and find that, in languages with relatively simple morphology, orthographic regularities on their own allow inflection models to achieve respectable accuracy. Combined orthographic and semantic regularities alleviate difficulties with particularly complex morpho-phonological systems. Our results suggest that hand-crafting many tagged examples might be an unnecessary effort. However, more work is needed in order to address rarely used forms.

(2 min) Recent work on using natural language to specify commands to robots has grounded that language to LTL. However, mapping natural language task specifications to LTL task specifications using language models require probability distributions over finite vocabulary. Existing state-of-the-art methods have extended this finite vocabulary to include unseen terms from the input sequence to improve output generalization. However, novel out-of-vocabulary atomic propositions cannot be generated using these methods. To overcome this, we introduce an intermediate contextual query representation which can be learned from single positive task specification examples, associating a contextual query with an LTL template. We demonstrate that this intermediate representation allows for generalization over unseen object references, assuming accurate groundings are available. We compare our method of mapping natural language task specifications to intermediate contextual queries against state-of-the-art CopyNet models capable of translating natural language to LTL, by evaluating whether correct LTL for manipulation and navigation task specifications can be output, and show that our method outperforms the CopyNet model on unseen object references. We demonstrate that the grounded LTL our method outputs can be used for planning in a simulated OO-MDP environment. Finally, we discuss some common failure modes encountered when translating natural language task specifications to grounded LTL.

(2 min) Differentially Private (DP) learning has seen limited success for building large deep learning models of text, and attempts at straightforwardly applying Differentially Private Stochastic Gradient Descent (DP-SGD) to NLP tasks have resulted in large performance drops and high computational overhead. We show that this performance drop can be mitigated with (1) the use of large pretrained models; (2) hyperparameters that suit DP optimization; and (3) fine-tuning objectives aligned with the pretraining procedure. With these factors set right, we obtain private NLP models that outperform state-of-the-art private training approaches and strong non-private baselines -- by directly fine-tuning pretrained models with DP optimization on moderately-sized corpora. To address the computational challenge of running DP-SGD with large Transformers, we propose a memory saving technique that allows clipping in DP-SGD to run without instantiating per-example gradients for any layer in the model. The technique enables privately training Transformers with almost the same memory cost as non-private training at a modest run-time overhead. Contrary to conventional wisdom that DP optimization fails at learning high-dimensional models (due to noise that scales with dimension) empirical results reveal that private learning with pretrained models tends to not suffer from dimension-dependent performance degradation.

(2 min) AI technologies for Natural Languages have made tremendous progress recently. However, commensurate progress has not been made on Sign Languages, in particular, in recognizing signs as individual words or as complete sentences. We introduce OpenHands, a library where we take four key ideas from the NLP community for low-resource languages and apply them to sign languages for word-level recognition. First, we propose using pose extracted through pretrained models as the standard modality of data to reduce training time and enable efficient inference, and we release standardized pose datasets for 6 different sign languages - American, Argentinian, Chinese, Greek, Indian, and Turkish. Second, we train and release checkpoints of 4 pose-based isolated sign language recognition models across all 6 languages, providing baselines and ready checkpoints for deployment. Third, to address the lack of labelled data, we propose self-supervised pretraining on unlabelled data. We curate and release the largest pose-based pretraining dataset on Indian Sign Language (Indian-SL). Fourth, we compare different pretraining strategies and for the first time establish that pretraining is effective for sign language recognition by demonstrating (a) improved fine-tuning performance especially in low-resource settings, and (b) high crosslingual transfer from Indian-SL to few other sign languages. We open-source all models and datasets in OpenHands with a hope that it makes research in sign languages more accessible, available here at https://github.com/AI4Bharat/OpenHands .

(2 min) SOTA coreference resolution produces increasingly impressive scores on the OntoNotes benchmark. However lack of comparable data following the same scheme for more genres makes it difficult to evaluate generalizability to open domain data. Zhu et al. (2021) introduced the creation of the OntoGUM corpus for evaluating geralizability of the latest neural LM-based end-to-end systems. This paper covers details of the mapping process which is a set of deterministic rules applied to the rich syntactic and discourse annotations manually annotated in the GUM corpus. Out-of-domain evaluation across 12 genres shows nearly 15-20% degradation for both deterministic and deep learning systems, indicating a lack of generalizability or covert overfitting in existing coreference resolution models.

(2 min) Continuous speech separation using a microphone array was shown to be promising in dealing with the speech overlap problem in natural conversation transcription. This paper proposes VarArray, an array-geometry-agnostic speech separation neural network model. The proposed model is applicable to any number of microphones without retraining while leveraging the nonlinear correlation between the input channels. The proposed method adapts different elements that were proposed before separately, including transform-average-concatenate, conformer speech separation, and inter-channel phase differences, and combines them in an efficient and cohesive way. Large-scale evaluation was performed with two real meeting transcription tasks by using a fully developed transcription system requiring no prior knowledge such as reference segmentations, which allowed us to measure the impact that the continuous speech separation system could have in realistic settings. The proposed model outperformed a previous approach to array-geometry-agnostic modeling for all of the geometry configurations considered, achieving asclite-based speaker-agnostic word error rates of 17.5% and 20.4% for the AMI development and evaluation sets, respectively, in the end-to-end setting using no ground-truth segmentations.

(2 min) This paper measures the impact of increased exposure on whether learned construction grammars converge onto shared representations when trained on data from different registers. Register influences the frequency of constructions, with some structures common in formal but not informal usage. We expect that a grammar induction algorithm exposed to different registers will acquire different constructions. To what degree does increased exposure lead to the convergence of register-specific grammars? The experiments in this paper simulate language learning in 12 languages (half Germanic and half Romance) with corpora representing three registers (Twitter, Wikipedia, Web). These simulations are repeated with increasing amounts of exposure, from 100k to 2 million words, to measure the impact of exposure on the convergence of grammars. The results show that increased exposure does lead to converging grammars across all languages. In addition, a shared core of register-universal constructions remains constant across increasing amounts of exposure.

(2 min) Natural language understanding tasks such as open-domain question answering often require retrieving and assimilating factual information from multiple sources. We propose to address this problem by integrating a semi-parametric representation of a large text corpus into a Transformer model as a source of factual knowledge. Specifically, our method represents knowledge with `mention memory', a table of dense vector representations of every entity mention in a corpus. The proposed model - TOME - is a Transformer that accesses the information through internal memory layers in which each entity mention in the input passage attends to the mention memory. This approach enables synthesis of and reasoning over many disparate sources of information within a single Transformer model. In experiments using a memory of 150 million Wikipedia mentions, TOME achieves strong performance on several open-domain knowledge-intensive tasks, including the claim verification benchmarks HoVer and FEVER and several entity-based QA benchmarks. We also show that the model learns to attend to informative mentions without any direct supervision. Finally we demonstrate that the model can generalize to new unseen entities by updating the memory without retraining.

(2 min) Self-supervised learning (SSL) is a long-standing goal for speech processing, since it utilizes large-scale unlabeled data and avoids extensive human labeling. Recent years witness great successes in applying self-supervised learning in speech recognition, while limited exploration was attempted in applying SSL for modeling speaker characteristics. In this paper, we aim to improve the existing SSL framework for speaker representation learning. Two methods are introduced for enhancing the unsupervised speaker information extraction. First, we apply the multi-task learning to the current SSL framework, where we integrate the utterance-wise contrastive loss with the SSL objective function. Second, for better speaker discrimination, we propose an utterance mixing strategy for data augmentation, where additional overlapped utterances are created unsupervisely and incorporate during training. We integrate the proposed methods into the HuBERT framework. Experiment results on SUPERB benchmark show that the proposed system achieves state-of-the-art performance in universal representation learning, especially for speaker identification oriented tasks. An ablation study is performed verifying the efficacy of each proposed method. Finally, we scale up training dataset to 94 thousand hours public audio data and achieve further performance improvement in all SUPERB tasks.

(2 min) There are two cases describing how a classifier processes input text, namely, misclassification and correct classification. In terms of misclassified texts, a classifier handles the texts with both incorrect predictions and adversarial texts, which are generated to fool the classifier, which is called a victim. Both types are misunderstood by the victim, but they can still be recognized by other classifiers. This induces large gaps in predicted probabilities between the victim and the other classifiers. In contrast, text correctly classified by the victim is often successfully predicted by the others and induces small gaps. In this paper, we propose an ensemble model based on similarity estimation of predicted probabilities (SEPP) to exploit the large gaps in the misclassified predictions in contrast to small gaps in the correct classification. SEPP then corrects the incorrect predictions of the misclassified texts. We demonstrate the resilience of SEPP in defending and detecting adversarial texts through different types of victim classifiers, classification tasks, and adversarial attacks.

(2 min) A popular approach to decompose the neural bases of language consists in correlating, across individuals, the brain responses to different stimuli (e.g. regular speech versus scrambled words, sentences, or paragraphs). Although successful, this `model-free' approach necessitates the acquisition of a large and costly set of neuroimaging data. Here, we show that a model-based approach can reach equivalent results within subjects exposed to natural stimuli. We capitalize on the recently-discovered similarities between deep language models and the human brain to compute the mapping between i) the brain responses to regular speech and ii) the activations of deep language models elicited by modified stimuli (e.g. scrambled words, sentences, or paragraphs). Our model-based approach successfully replicates the seminal study of Lerner et al. (2011), which revealed the hierarchy of language areas by comparing the functional-magnetic resonance imaging (fMRI) of seven subjects listening to 7min of both regular and scrambled narratives. We further extend and precise these results to the brain signals of 305 individuals listening to 4.1 hours of narrated stories. Overall, this study paves the way for efficient and flexible analyses of the brain bases of language.

(2 min) Recent work like GPT-3 has demonstrated excellent performance of Zero-Shot and Few-Shot learning on many natural language processing (NLP) tasks by scaling up model size, dataset size and the amount of computation. However, training a model like GPT-3 requires huge amount of computational resources which makes it challengeable to researchers. In this work, we propose a method that incorporates large-scale distributed training performance into model architecture design. With this method, Yuan 1.0, the current largest singleton language model with 245B parameters, achieves excellent performance on thousands GPUs during training, and the state-of-the-art results on NLP tasks. A data processing method is designed to efficiently filter massive amount of raw data. The current largest high-quality Chinese corpus with 5TB high quality texts is built based on this method. In addition, a calibration and label expansion method is proposed to improve the Zero-Shot and Few-Shot performance, and steady improvement is observed on the accuracy of various tasks. Yuan 1.0 presents strong capacity of natural language generation, and the generated articles are difficult to distinguish from the human-written ones.

(2 min) The task of generating rich and fluent narratives that aptly describe the characteristics, trends, and anomalies of time-series data is invaluable to the sciences (geology, meteorology, epidemiology) or finance (trades, stocks, or sales and inventory). The efforts for time-series narration hitherto are domain-specific and use predefined templates that offer consistency but lead to mechanical narratives. We present TCube (Time-series-to-text), a domain-agnostic neural framework for time-series narration, that couples the representation of essential time-series elements in the form of a dense knowledge graph and the translation of said knowledge graph into rich and fluent narratives through the transfer-learning capabilities of PLMs (Pre-trained Language Models). TCube's design primarily addresses the challenge that lies in building a neural framework in the complete paucity of annotated training data for time-series. The design incorporates knowledge graphs as an intermediary for the representation of essential time-series elements which can be linearized for textual translation. To the best of our knowledge, TCube is the first investigation of the use of neural strategies for time-series narration. Through extensive evaluations, we show that TCube can improve the lexical diversity of the generated narratives by up to 65.38% while still maintaining grammatical integrity. The practicality and deployability of TCube is further validated through an expert review (n=21) where 76.2% of participating experts wary of auto-generated narratives favored TCube as a deployable system for time-series narration due to its richer narratives. Our code-base, models, and datasets, with detailed instructions for reproducibility is publicly hosted at https://github.com/Mandar-Sharma/TCube.

(2 min) In this paper, we investigate the knowledge distillation (KD) strategy for object detection and propose an effective framework applicable to both homogeneous and heterogeneous student-teacher pairs. The conventional feature imitation paradigm introduces imitation masks to focus on informative foreground areas while excluding the background noises. However, we find that those methods fail to fully utilize the semantic information in all feature pyramid levels, which leads to inefficiency for knowledge distillation between FPN-based detectors. To this end, we propose a novel semantic-guided feature imitation technique, which automatically performs soft matching between feature pairs across all pyramid levels to provide the optimal guidance to the student. To push the envelop even further, we introduce contrastive distillation to effectively capture the information encoded in the relationship between different feature regions. Finally, we propose a generalized detection KD pipeline, which is capable of distilling both homogeneous and heterogeneous detector pairs. Our method consistently outperforms the existing detection KD techniques, and works when (1) components in the framework are used separately and in conjunction; (2) for both homogeneous and heterogenous student-teacher pairs and (3) on multiple detection benchmarks. With a powerful X101-FasterRCNN-Instaboost detector as the teacher, R50-FasterRCNN reaches 44.0% AP, R50-RetinaNet reaches 43.3% AP and R50-FCOS reaches 43.1% AP on COCO dataset.

(0 min) Action recognition is a crucial task for video understanding. In this paper, we present AutoVideo, a Python system for automated video action recognition. It currently supports seven action recognition algorithms and various pre-processing modules. Unlike the existing libraries that only provide model zoos, AutoVideo is built with the standard pipeline language. The basic building block is primitive, which wraps a pre-processing module or an algorithm with some hyperparameters. AutoVideo is highly modular and extendable. It can be easily combined with AutoML searchers. The pipeline language is quite general so that we can easily enrich AutoVideo with algorithms for various other video-related tasks in the future. AutoVideo is released under MIT license at https://github.com/datamllab/autovideo

(2 min) Generative diffusion processes are an emerging and effective tool for image and speech generation. In the existing methods, the underlying noise distribution of the diffusion process is Gaussian noise. However, fitting distributions with more degrees of freedom could improve the performance of such generative models. In this work, we investigate other types of noise distribution for the diffusion process. Specifically, we introduce the Denoising Diffusion Gamma Model (DDGM) and show that noise from Gamma distribution provides improved results for image and speech generation. Our approach preserves the ability to efficiently sample state in the training diffusion process while using Gamma noise.

(0 min) Depth information is essential for on-board perception in autonomous driving and driver assistance. Monocular depth estimation (MDE) is very appealing since it allows for appearance and depth being on direct pixelwise correspondence without further calibration. Best MDE models are based on Convolutional Neural Networks (CNNs) trained in a supervised manner, i.e., assuming pixelwise ground truth (GT). Usually, this GT is acquired at training time through a calibrated multi-modal suite of sensors. However, also using only a monocular system at training time is cheaper and more scalable. This is possible by relying on structure-from-motion (SfM) principles to generate self-supervision. Nevertheless, problems of camouflaged objects, visibility changes, static-camera intervals, textureless areas, and scale ambiguity, diminish the usefulness of such self-supervision. In this paper, we perform monocular depth estimation by virtual-world supervision (MonoDEVS) and real-world SfM self-supervision. We compensate the SfM self-supervision limitations by leveraging virtual-world images with accurate semantic and depth supervision and addressing the virtual-to-real domain gap. Our MonoDEVSNet outperforms previous MDE CNNs trained on monocular and even stereo sequences.

(0 min) Optical Character Recognition (OCR) systems have been widely used in various of application scenarios. Designing an OCR system is still a challenging task. In previous work, we proposed a practical ultra lightweight OCR system (PP-OCR) to balance the accuracy against the efficiency. In order to improve the accuracy of PP-OCR and keep high efficiency, in this paper, we propose a more robust OCR system, i.e. PP-OCRv2. We introduce bag of tricks to train a better text detector and a better text recognizer, which include Collaborative Mutual Learning (CML), CopyPaste, Lightweight CPUNetwork (LCNet), Unified-Deep Mutual Learning (U-DML) and Enhanced CTCLoss. Experiments on real data show that the precision of PP-OCRv2 is 7% higher than PP-OCR under the same inference cost. It is also comparable to the server models of the PP-OCR which uses ResNet series as backbones. All of the above mentioned models are open-sourced and the code is available in the GitHub repository PaddleOCR which is powered by PaddlePaddle.

(0 min) Self-supervised representation learning is able to learn semantically meaningful features; however, much of its recent success relies on multiple crops of an image with very few objects. Instead of learning view-invariant representation from simple images, humans learn representations in a complex world with changing scenes by observing object movement, deformation, pose variation, and ego motion. Motivated by this ability, we present a new self-supervised learning representation framework that can be directly deployed on a video stream of complex scenes with many moving objects. Our framework features a simple flow equivariance objective that encourages the network to predict the features of another frame by applying a flow transformation to the features of the current frame. Our representations, learned from high-resolution raw video, can be readily used for downstream tasks on static images. Readout experiments on challenging semantic segmentation, instance segmentation, and object detection benchmarks show that we are able to outperform representations obtained from previous state-of-the-art methods including SimCLR and BYOL.

(0 min) Fine-grained human action recognition is a core research topic in computer vision. Inspired by the recently proposed hierarchy representation of fine-grained actions in FineGym and SlowFast network for action recognition, we propose a novel multi-task network which exploits the FineGym hierarchy representation to achieve effective joint learning and prediction for fine-grained human action recognition. The multi-task network consists of three pathways of SlowOnly networks with gradually increased frame rates for events, sets and elements of fine-grained actions, followed by our proposed integration layers for joint learning and prediction. It is a two-stage approach, where it first learns deep feature representation at each hierarchical level, and is followed by feature encoding and fusion for multi-task learning. Our empirical results on the FineGym dataset achieve a new state-of-the-art performance, with 91.80% Top-1 accuracy and 88.46% mean accuracy for element actions, which are 3.40% and 7.26% higher than the previous best results.

(0 min) Image deraining is a new challenging problem in applications of autonomous vehicles. In a bad weather condition of heavy rainfall, raindrops, mainly hitting the vehicle's windshield, can significantly reduce observation ability even though the windshield wipers might be able to remove part of it. Moreover, rain flows spreading over the windshield can yield the physical effect of refraction, which seriously impede the sightline or undermine the machine learning system equipped in the vehicle. In this paper, we propose a new multi-stage multi-task recurrent generative adversarial network (M2GAN) to deal with challenging problems of raindrops hitting the car's windshield. This method is also applicable for removing raindrops appearing on a glass window or lens. M2GAN is a multi-stage multi-task generative adversarial network that can utilize prior high-level information, such as semantic segmentation, to boost deraining performance. To demonstrate M2GAN, we introduce the first real-world dataset for rain removal on autonomous vehicles. The experimental results show that our proposed method is superior to other state-of-the-art approaches of deraining raindrops in respect of quantitative metrics and visual quality. M2GAN is considered the first method to deal with challenging problems of real-world rains under unconstrained environments such as autonomous vehicles.

(0 min) The existing image embedding networks are basically vulnerable to malicious attacks such as JPEG compression and noise adding, not applicable for real-world copyright protection tasks. To solve this problem, we introduce a generative deep network based method for hiding images into images while assuring high-quality extraction from the destructive synthesized images. An embedding network is sequentially concatenated with an attack layer, a decoupling network and an image extraction network. The addition of decoupling network learns to extract the embedded watermark from the attacked image. We also pinpoint the weaknesses of the adversarial training for robustness in previous works and build our improved real-world attack simulator. Experimental results demonstrate the superiority of the proposed method against typical digital attacks by a large margin, as well as the performance boost of the recovered images with the aid of progressive recovery strategy. Besides, we are the first to robustly hide three secret images.

(0 min) The simultaneous recognition of multiple objects in one image remains a challenging task, spanning multiple events in the recognition field such as various object scales, inconsistent appearances, and confused inter-class relationships. Recent research efforts mainly resort to the statistic label co-occurrences and linguistic word embedding to enhance the unclear semantics. Different from these researches, in this paper, we propose a novel Transformer-based Dual Relation learning framework, constructing complementary relationships by exploring two aspects of correlation, i.e., structural relation graph and semantic relation graph. The structural relation graph aims to capture long-range correlations from object context, by developing a cross-scale transformer-based architecture. The semantic graph dynamically models the semantic meanings of image objects with explicit semantic-aware constraints. In addition, we also incorporate the learnt structural relationship into the semantic graph, constructing a joint relation graph for robust representations. With the collaborative learning of these two effective relation graphs, our approach achieves new state-of-the-art on two popular multi-label recognition benchmarks, i.e., MS-COCO and VOC 2007 dataset.

(0 min) This paper unifies the multi-focus images fusion (MFIF) and blind super resolution (SR) problems as the multi-focus image super resolution fusion (MFISRF) task, and proposes a novel unified dataset-free unsupervised framework named deep fusion prior (DFP) to address such MFISRF task. DFP consists of SKIPnet network, DoubleReblur focus measurement tactic, decision embedding module and loss functions. In particular, DFP can obtain MFISRF only from two low-resolution inputs without any extent dataset; SKIPnet implementing unsupervised learning via deep image prior is an end-to-end generated network acting as the engine of DFP; DoubleReblur is used to determine the primary decision map without learning but based on estimated PSF and Gaussian kernels convolution; decision embedding module optimizes the decision map via learning; and DFP losses composed of content loss, joint gradient loss and gradient limit loss can obtain high-quality MFISRF results robustly. Experiments have proved that our proposed DFP approaches and even outperforms those state-of-art MFIF and SR method combinations. Additionally, DFP is a general framework, thus its networks and focus measurement tactics can be continuously updated to further improve the MFISRF performance. DFP codes are open source and will be available soon at this http URL

(0 min) The energy consumption of deep learning models is increasing at a breathtaking rate, which raises concerns due to potential negative effects on carbon neutrality in the context of global warming and climate change. With the progress of efficient deep learning techniques, e.g., model compression, researchers can obtain efficient models with fewer parameters and smaller latency. However, most of the existing efficient deep learning methods do not explicitly consider energy consumption as a key performance indicator. Furthermore, existing methods mostly focus on the inference costs of the resulting efficient models, but neglect the notable energy consumption throughout the entire life cycle of the algorithm. In this paper, we present the first large-scale energy consumption benchmark for efficient computer vision models, where a new metric is proposed to explicitly evaluate the full-cycle energy consumption under different model usage intensity. The benchmark can provide insights for low carbon emission when selecting efficient deep learning algorithms in different model usage scenarios.

(0 min) Image deraining is a new challenging problem in real-world applications, such as autonomous vehicles. In a bad weather condition of heavy rainfall, raindrops, mainly hitting glasses or windshields, can significantly reduce observation ability. Moreover, raindrops spreading over the glass can yield refraction's physical effect, which seriously impedes the sightline or undermine machine learning systems. In this paper, we propose an enhanced generative adversarial network to deal with the challenging problems of raindrops. UnfairGAN is an enhanced generative adversarial network that can utilize prior high-level information, such as edges and rain estimation, to boost deraining performance. To demonstrate UnfairGAN, we introduce a large dataset for training deep learning models of rain removal. The experimental results show that our proposed method is superior to other state-of-the-art approaches of deraining raindrops regarding quantitative metrics and visual quality.

(0 min) Multi-modal imaging is a key healthcare technology that is often underutilized due to costs associated with multiple separate scans. This limitation yields the need for synthesis of unacquired modalities from the subset of available modalities. In recent years, generative adversarial network (GAN) models with superior depiction of structural details have been established as state-of-the-art in numerous medical image synthesis tasks. GANs are characteristically based on convolutional neural network (CNN) backbones that perform local processing with compact filters. This inductive bias in turn compromises learning of contextual features. Here, we propose a novel generative adversarial approach for medical image synthesis, ResViT, to combine local precision of convolution operators with contextual sensitivity of vision transformers. ResViT employs a central bottleneck comprising novel aggregated residual transformer (ART) blocks that synergistically combine convolutional and transformer modules. Comprehensive demonstrations are performed for synthesizing missing sequences in multi-contrast MRI, and CT images from MRI. Our results indicate superiority of ResViT against competing methods in terms of qualitative observations and quantitative metrics.

(0 min) We present LaLaLoc to localise in environments without the need for prior visitation, and in a manner that is robust to large changes in scene appearance, such as a full rearrangement of furniture. Specifically, LaLaLoc performs localisation through latent representations of room layout. LaLaLoc learns a rich embedding space shared between RGB panoramas and layouts inferred from a known floor plan that encodes the structural similarity between locations. Further, LaLaLoc introduces direct, cross-modal pose optimisation in its latent space. Thus, LaLaLoc enables fine-grained pose estimation in a scene without the need for prior visitation, as well as being robust to dynamics, such as a change in furniture configuration. We show that in a domestic environment LaLaLoc is able to accurately localise a single RGB panorama image to within 8.3cm, given only a floor plan as a prior.

(0 min) Since the rise of deep learning, many computer vision tasks have seen significant advancements. However, the downside of deep learning is that it is very data-hungry. Especially for segmentation problems, training a deep neural net requires dense supervision in the form of pixel-perfect image labels, which are very costly. In this paper, we present a new loss function to train a segmentation network with only a small subset of pixel-perfect labels, but take the advantage of weakly-annotated training samples in the form of cheap bounding-box labels. Unlike recent works which make use of box-to-mask proposal generators, our loss trains the network to learn a label uncertainty within the bounding-box, which can be leveraged to perform online bootstrapping (i.e. transforming the boxes to segmentation masks), while training the network. We evaluated our method on binary segmentation tasks, as well as a multi-class segmentation task (CityScapes vehicles and persons). We trained each task on a dataset comprised of only 18% pixel-perfect and 82% bounding-box labels, and compared the results to a baseline model trained on a completely pixel-perfect dataset. For the binary segmentation tasks, our method achieves an IoU score which is ~98.33% as good as our baseline model, while for the multi-class task, our method is 97.12% as good as our baseline model (77.5 vs. 79.8 mIoU).

(0 min) The prototypical network is a prototype classifier based on meta-learning and is widely used for few-shot learning because it classifies unseen examples by constructing class-specific prototypes without adjusting hyper-parameters during meta-testing. Interestingly, recent research has attracted a lot of attention, showing that a linear classifier with fine-tuning, which does not use a meta-learning algorithm, performs comparably with the prototypical network. However, fine-tuning requires additional hyper-parameters when adapting a model to a new environment. In addition, although the purpose of few-shot learning is to enable the model to quickly adapt to a new environment, fine-tuning needs to be applied every time a new class appears, making fast adaptation difficult. In this paper, we analyze how a prototype classifier works equally well without fine-tuning and meta-learning. We experimentally found that directly using the feature vector extracted using standard pre-trained models to construct a prototype classifier in meta-testing does not perform as well as the prototypical network and linear classifiers with fine-tuning and feature vectors of pre-trained models. Thus, we derive a novel generalization bound for the prototypical network and show that focusing on the variance of the norm of a feature vector can improve performance. We experimentally investigated several normalization methods for minimizing the variance of the norm and found that the same performance can be obtained by using the L2 normalization and embedding space transformation without fine-tuning or meta-learning.

(0 min) Motion and dynamic environments, especially under challenging lighting conditions, are still an open issue for robust robotic applications. In this paper, we propose an end-to-end pipeline for real-time, low latency, 6 degrees-of-freedom pose estimation of fiducial markers. Instead of achieving a pose estimation through a conventional frame-based approach, we employ the high-speed abilities of event-based sensors to directly refine the spatial transformation, using consecutive events. Furthermore, we introduce a novel two-way verification process for detecting tracking errors by backtracking the estimated pose, allowing us to evaluate the quality of our tracking. This approach allows us to achieve pose estimation at a rate up to 156~kHz, while only relying on CPU resources. The average end-to-end latency of our method is 3~ms. Experimental results demonstrate outstanding potential for robotic tasks, such as visual servoing in fast action-perception loops.

(0 min) DETR is a recently proposed Transformer-based method which views object detection as a set prediction problem and achieves state-of-the-art performance but demands extra-long training time to converge. In this paper, we investigate the causes of the optimization difficulty in the training of DETR. Our examinations reveal several factors contributing to the slow convergence of DETR, primarily the issues with the Hungarian loss and the Transformer cross-attention mechanism. To overcome these issues we propose two solutions, namely, TSP-FCOS (Transformer-based Set Prediction with FCOS) and TSP-RCNN (Transformer-based Set Prediction with RCNN). Experimental results show that the proposed methods not only converge much faster than the original DETR, but also significantly outperform DETR and other baselines in terms of detection accuracy.

(0 min) 3D object detection with LiDAR point clouds plays an important role in autonomous driving perception module that requires high speed, stability and accuracy. However, the existing point-based methods are challenging to reach the speed requirements because of too many raw points, and the voxel-based methods are unable to ensure stable speed because of the 3D sparse convolution. In contrast, the 2D grid-based methods, such as PointPillar, can easily achieve a stable and efficient speed based on simple 2D convolution, but it is hard to get the competitive accuracy limited by the coarse-grained point clouds representation. So we propose an improved pillar with fine-grained feature based on PointPillar that can significantly improve detection accuracy. It consists of two modules, including height-aware sub-pillar and sparsity-based tiny-pillar, which get fine-grained representation respectively in the vertical and horizontal direction of 3D space. For height-aware sub-pillar, we introduce a height position encoding to keep height information of each sub-pillar during projecting to a 2D pseudo image. For sparsity-based tiny-pillar, we introduce sparsity-based CNN backbone stacked by dense feature and sparse attention module to extract feature with larger receptive field efficiently. Experimental results show that our proposed method significantly outperforms previous state-of-the-art 3D detection methods on the Waymo Open Dataset. The related code will be released to facilitate the academic and industrial study.

(0 min) Video-text retrieval has many real-world applications such as media analytics, surveillance, and robotics. This paper presents the 1st place solution to the video retrieval track of the ICCV VALUE Challenge 2021. We present a simple yet effective approach to jointly tackle two video-text retrieval tasks (video retrieval and video corpus moment retrieval) by leveraging the model trained only on the video retrieval task. In addition, we create an ensemble model that achieves the new state-of-the-art performance on all four datasets (TVr, How2r, YouCook2r, and VATEXr) presented in the VALUE Challenge.

(0 min) Depending on the application, radiological diagnoses can be associated with high inter- and intra-rater variabilities. Most computer-aided diagnosis (CAD) solutions treat such data as incontrovertible, exposing learning algorithms to considerable and possibly contradictory label noise and biases. Thus, managing subjectivity in labels is a fundamental problem in medical imaging analysis. To address this challenge, we introduce auto-decoded deep latent embeddings (ADDLE), which explicitly models the tendencies of each rater using an auto-decoder framework. After a simple linear transformation, the latent variables can be injected into any backbone at any and multiple points, allowing the model to account for rater-specific effects on the diagnosis. Importantly, ADDLE does not expect multiple raters per image in training, meaning it can readily learn from data mined from hospital archives. Moreover, the complexity of training ADDLE does not increase as more raters are added. During inference each rater can be simulated and a 'mean' or 'greedy' virtual rating can be produced. We test ADDLE on the problem of liver steatosis diagnosis from 2D ultrasound (US) by collecting 46 084 studies along with clinical US diagnoses originating from 65 different raters. We evaluated diagnostic performance using a separate dataset with gold-standard biopsy diagnoses. ADDLE can improve the partial areas under the curve (AUCs) for diagnosing severe steatosis by 10.5% over standard classifiers while outperforming other annotator-noise approaches, including those requiring 65 times the parameters.

(0 min) Sign language is commonly used by deaf or mute people to communicate but requires extensive effort to master. It is usually performed with the fast yet delicate movement of hand gestures, body posture, and even facial expressions. Current Sign Language Recognition (SLR) methods usually extract features via deep neural networks and suffer overfitting due to limited and noisy data. Recently, skeleton-based action recognition has attracted increasing attention due to its subject-invariant and background-invariant nature, whereas skeleton-based SLR is still under exploration due to the lack of hand annotations. Some researchers have tried to use off-line hand pose trackers to obtain hand keypoints and aid in recognizing sign language via recurrent neural networks. Nevertheless, none of them outperforms RGB-based approaches yet. To this end, we propose a novel Skeleton Aware Multi-modal Framework with a Global Ensemble Model (GEM) for isolated SLR (SAM-SLR-v2) to learn and fuse multi-modal feature representations towards a higher recognition rate. Specifically, we propose a Sign Language Graph Convolution Network (SL-GCN) to model the embedded dynamics of skeleton keypoints and a Separable Spatial-Temporal Convolution Network (SSTCN) to exploit skeleton features. The skeleton-based predictions are fused with other RGB and depth based modalities by the proposed late-fusion GEM to provide global information and make a faithful SLR prediction. Experiments on three isolated SLR datasets demonstrate that our proposed SAM-SLR-v2 framework is exceedingly effective and achieves state-of-the-art performance with significant margins. Our code will be available at https://github.com/jackyjsy/SAM-SLR-v2

(0 min) Recent work in cognitive reasoning and computer vision has engendered an increasing popularity for the Violation-of-Expectation (VoE) paradigm in synthetic datasets. Inspired by work in infant psychology, researchers have started evaluating a model's ability to discriminate between expected and surprising scenes as a sign of its reasoning ability. Existing VoE-based 3D datasets in physical reasoning only provide vision data. However, current cognitive models of physical reasoning by psychologists reveal infants create high-level abstract representations of objects and interactions. Capitalizing on this knowledge, we propose AVoE: a synthetic 3D VoE-based dataset that presents stimuli from multiple novel sub-categories for five event categories of physical reasoning. Compared to existing work, AVoE is armed with ground-truth labels of abstract features and rules augmented to vision data, paving the way for high-level symbolic predictions in physical reasoning tasks.

(0 min) In this paper, we study the spectral properties of re-parameterized light field. Following previous studies of the light field spectrum, which notably provided sampling guidelines, we focus on the two plane parameterization of the light field. However, we introduce additional flexibility by allowing the image plane to be tilted and not only parallel. A formal theoretical analysis is first presented, which shows that more flexible sampling guidelines (i.e. wider camera baselines) can be used to sample the light field when adapting the image plane orientation to the scene geometry. We then present our simulations and results to support these theoretical findings. While the work introduced in this paper is mostly theoretical, we believe these new findings open exciting avenues for more practical application of light fields, such as view synthesis or compact representation.

(0 min) Most video super-resolution methods focus on restoring high-resolution video frames from low-resolution videos without taking into account compression. However, most videos on the web or mobile devices are compressed, and the compression can be severe when the bandwidth is limited. In this paper, we propose a new compression-informed video super-resolution model to restore high-resolution content without introducing artifacts caused by compression. The proposed model consists of three modules for video super-resolution: bi-directional recurrent warping, detail-preserving flow estimation, and Laplacian enhancement. All these three modules are used to deal with compression properties such as the location of the intra-frames in the input and smoothness in the output frames. For thorough performance evaluation, we conducted extensive experiments on standard datasets with a wide range of compression rates, covering many real video use cases. We showed that our method not only recovers high-resolution content on uncompressed frames from the widely-used benchmark datasets, but also achieves state-of-the-art performance in super-resolving compressed videos based on numerous quantitative metrics. We also evaluated the proposed method by simulating streaming from YouTube to demonstrate its effectiveness and robustness. The source codes and trained models are available at https://github.com/google-research/google-research/tree/master/comisr.

(0 min) This paper focuses on tackling the problem of temporal language localization in videos, which aims to identify the start and end points of a moment described by a natural language sentence in an untrimmed video. However, it is non-trivial since it requires not only the comprehensive understanding of the video and sentence query, but also the accurate semantic correspondence capture between them. Existing efforts are mainly centered on exploring the sequential relation among video clips and query words to reason the video and sentence query, neglecting the other intra-modal relations (e.g., semantic similarity among video clips and syntactic dependency among the query words). Towards this end, in this work, we propose a Multi-modal Interaction Graph Convolutional Network (MIGCN), which jointly explores the complex intra-modal relations and inter-modal interactions residing in the video and sentence query to facilitate the understanding and semantic correspondence capture of the video and sentence query. In addition, we devise an adaptive context-aware localization method, where the context information is taken into the candidate moments and the multi-scale fully connected layers are designed to rank and adjust the boundary of the generated coarse candidate moments with different lengths. Extensive experiments on Charades-STA and ActivityNet datasets demonstrate the promising performance and superior efficiency of our model.

(0 min) Medical image analysis continues to hold interesting challenges given the subtle characteristics of certain diseases and the significant overlap in appearance between diseases. In this work, we explore the concept of self-attention for tackling such subtleties in and between diseases. To this end, we introduce MEDUSA, a multi-scale encoder-decoder self-attention mechanism tailored for medical image analysis. While self-attention deep convolutional neural network architectures in existing literature center around the notion of multiple isolated lightweight attention mechanisms with limited individual capacities being incorporated at different points in the network architecture, MEDUSA takes a significant departure from this notion by possessing a single, unified self-attention mechanism with significantly higher capacity with multiple attention heads feeding into different scales in the network architecture. To the best of the authors' knowledge, this is the first "single body, multi-scale heads" realization of self-attention and enables explicit global context amongst selective attention at different levels of representational abstractions while still enabling differing local attention context at individual levels of abstractions. With MEDUSA, we obtain state-of-the-art performance on multiple challenging medical image analysis benchmarks including COVIDx, RSNA RICORD, and RSNA Pneumonia Challenge when compared to previous work. Our MEDUSA model is publicly available.

(0 min) Semantic concepts are frequently defined by combinations of underlying attributes. As mappings from attributes to classes are often simple, attribute-based representations facilitate novel concept learning with zero or few examples. A significant limitation of existing attribute-based learning paradigms, such as zero-shot learning, is that the attributes are assumed to be known and fixed. In this work we study the rapid learning of attributes that were not previously labeled. Compared to standard few-shot learning of semantic classes, in which novel classes may be defined by attributes that were relevant at training time, learning new attributes imposes a stiffer challenge. We found that supervised learning with training attributes does not generalize well to new test attributes, whereas self-supervised pre-training brings significant improvement. We further experimented with random splits of the attribute space and found that predictability of test attributes provides an informative estimate of a model's generalization ability.

(0 min) It is well noted that coordinate based MLPs benefit -- in terms of preserving high-frequency information -- through the encoding of coordinate positions as an array of Fourier features. Hitherto, the rationale for the effectiveness of these positional encodings has been solely studied through a Fourier lens. In this paper, we strive to broaden this understanding by showing that alternative non-Fourier embedding functions can indeed be used for positional encoding. Moreover, we show that their performance is entirely determined by a trade-off between the stable rank of the embedded matrix and the distance preservation between embedded coordinates. We further establish that the now ubiquitous Fourier feature mapping of position is a special case that fulfills these conditions. Consequently, we present a more general theory to analyze positional encoding in terms of shifted basis functions. To this end, we develop the necessary theoretical formulae and empirically verify that our theoretical claims hold in practice. Codes available at https://github.com/osiriszjq/Rethinking-positional-encoding.

(0 min) Real world learning scenarios involve a nonstationary distribution of classes with sequential dependencies among the samples, in contrast to the standard machine learning formulation of drawing samples independently from a fixed, typically uniform distribution. Furthermore, real world interactions demand learning on-the-fly from few or no class labels. In this work, we propose an unsupervised model that simultaneously performs online visual representation learning and few-shot learning of new categories without relying on any class labels. Our model is a prototype-based memory network with a control component that determines when to form a new class prototype. We formulate it as an online Gaussian mixture model, where components are created online with only a single new example, and assignments do not have to be balanced, which permits an approximation to natural imbalanced distributions from uncurated raw data. Learning includes a contrastive loss that encourages different views of the same image to be assigned to the same prototype. The result is a mechanism that forms categorical representations of objects in nonstationary environments. Experiments show that our method can learn from an online stream of visual input data and is significantly better at category recognition compared to state-of-the-art self-supervised learning methods.

(0 min) Vision Transformers (ViTs) and MLPs signal further efforts on replacing hand-wired features or inductive biases with general-purpose neural architectures. Existing works empower the models by massive data, such as large-scale pre-training and/or repeated strong data augmentations, and still report optimization-related problems (e.g., sensitivity to initialization and learning rates). Hence, this paper investigates ViTs and MLP-Mixers from the lens of loss geometry, intending to improve the models' data efficiency at training and generalization at inference. Visualization and Hessian reveal extremely sharp local minima of converged models. By promoting smoothness with a recently proposed sharpness-aware optimizer, we substantially improve the accuracy and robustness of ViTs and MLP-Mixers on various tasks spanning supervised, adversarial, contrastive, and transfer learning (e.g., +5.3\% and +11.0\% top-1 accuracy on ImageNet for ViT-B/16 and Mixer-B/16, respectively, with the simple Inception-style preprocessing). We show that the improved smoothness attributes to sparser active neurons in the first few layers. The resultant ViTs outperform ResNets of similar size and throughput when trained from scratch on ImageNet without large-scale pre-training or strong data augmentations. They also possess more perceptive attention maps. Our model checkpoints are released at \url{https://github.com/google-research/vision_transformer}.

(0 min) Social distancing, an essential public health measure to limit the spread of contagious diseases, has gained significant attention since the outbreak of the COVID-19 pandemic. In this work, the problem of visual social distancing compliance assessment in busy public areas, with wide field-of-view cameras, is considered. A dataset of crowd scenes with people annotations under a bird's eye view (BEV) and ground truth for metric distances is introduced, and several measures for the evaluation of social distance detection systems are proposed. A multi-branch network, BEV-Net, is proposed to localize individuals in world coordinates and identify high-risk regions where social distancing is violated. BEV-Net combines detection of head and feet locations, camera pose estimation, a differentiable homography module to map image into BEV coordinates, and geometric reasoning to produce a BEV map of the people locations in the scene. Experiments on complex crowded scenes demonstrate the power of the approach and show superior performance over baselines derived from methods in the literature. Applications of interest for public health decision makers are finally discussed. Datasets, code and pretrained models are publicly available at GitHub.

(0 min) This paper presents a predictive model for estimating regularization parameters of diffeomorphic image registration. We introduce a novel framework that automatically determines the parameters controlling the smoothness of diffeomorphic transformations. Our method significantly reduces the effort of parameter tuning, which is time and labor-consuming. To achieve the goal, we develop a predictive model based on deep convolutional neural networks (CNN) that learns the mapping between pairwise images and the regularization parameter of image registration. In contrast to previous methods that estimate such parameters in a high-dimensional image space, our model is built in an efficient bandlimited space with much lower dimensions. We demonstrate the effectiveness of our model on both 2D synthetic data and 3D real brain images. Experimental results show that our model not only predicts appropriate regularization parameters for image registration, but also improving the network training in terms of time and memory efficiency.

(0 min) The ability to identify whether or not a test sample belongs to one of the semantic classes in a classifier's training set is critical to practical deployment of the model. This task is termed open-set recognition (OSR) and has received significant attention in recent years. In this paper, we first demonstrate that the ability of a classifier to make the 'none-of-above' decision is highly correlated with its accuracy on the closed-set classes. We find that this relationship holds across loss objectives and architectures, and further demonstrate the trend both on the standard OSR benchmarks as well as on a large-scale ImageNet evaluation. Second, we use this correlation to boost the performance of the cross-entropy OSR 'baseline' by improving its closed-set accuracy, and with this strong baseline achieve a new state-of-the-art on the most challenging OSR benchmark. Similarly, we boost the performance of the existing state-of-the-art method by improving its closed-set accuracy, but this does not surpass the strong baseline on the most challenging dataset. Our third contribution is to reappraise the datasets used for OSR evaluation, and construct new benchmarks which better respect the task of detecting semantic novelty, as opposed to low-level distributional shifts as tackled by neighbouring machine learning fields. In this new setting, we again demonstrate that there is negligible difference between the strong baseline and the existing state-of-the-art.

(0 min) Multi-modal reasoning systems rely on a pre-trained object detector to extract regions of interest from the image. However, this crucial module is typically used as a black box, trained independently of the downstream task and on a fixed vocabulary of objects and attributes. This makes it challenging for such systems to capture the long tail of visual concepts expressed in free form text. In this paper we propose MDETR, an end-to-end modulated detector that detects objects in an image conditioned on a raw text query, like a caption or a question. We use a transformer-based architecture to reason jointly over text and image by fusing the two modalities at an early stage of the model. We pre-train the network on 1.3M text-image pairs, mined from pre-existing multi-modal datasets having explicit alignment between phrases in text and objects in the image. We then fine-tune on several downstream tasks such as phrase grounding, referring expression comprehension and segmentation, achieving state-of-the-art results on popular benchmarks. We also investigate the utility of our model as an object detector on a given label set when fine-tuned in a few-shot setting. We show that our pre-training approach provides a way to handle the long tail of object categories which have very few labelled instances. Our approach can be easily extended for visual question answering, achieving competitive performance on GQA and CLEVR. The code and models are available at https://github.com/ashkamath/mdetr.

(0 min) Three-dimensional reconstruction of objects from shading information is a challenging task in computer vision. As most of the approaches facing the Photometric Stereo problem use simplified far-field assumptions, real-world scenarios have essentially more complex physical effects that need to be handled for accurately reconstructing the 3D shape. An increasing number of methods have been proposed to address the problem when point light sources are assumed to be nearby the target object. The proximity of the light sources complicates the modeling of the image formation as the light behaviour requires non-linear parameterisation to describe its propagation and attenuation. To understand the capability of the approaches dealing with this near-field scenario, the literature till now has used synthetically rendered photometric images or minimal and very customised real-world data. In order to fill the gap in evaluating near-field photometric stereo methods, we introduce LUCES the first real-world 'dataset for near-fieLd point light soUrCe photomEtric Stereo' of 14 objects of a varying of materials. A device counting 52 LEDs has been designed to lit each object positioned 10 to 30 centimeters away from the camera. Together with the raw images, in order to evaluate the 3D reconstructions, the dataset includes both normal and depth maps for comparing different features of the retrieved 3D geometry. Furthermore, we evaluate the performance of the latest near-field Photometric Stereo algorithms on the proposed dataset to assess the SOTA method with respect to actual close range effects and object materials.

(0 min) Hand pose estimation is a fundamental task in many human-robot interaction-related applications. However, previous approaches suffer from unsatisfying hand landmark predictions in real-world scenes and high computation burden. This paper proposes a fast and accurate framework for hand pose estimation, dubbed as "FastHand". Using a lightweight encoder-decoder network architecture, FastHand fulfills the requirements of practical applications running on embedded devices. The encoder consists of deep layers with a small number of parameters, while the decoder makes use of spatial location information to obtain more accurate results. The evaluation took place on two publicly available datasets demonstrating the improved performance of the proposed pipeline compared to other state-of-the-art approaches. FastHand offers high accuracy scores while reaching a speed of 25 frames per second on an NVIDIA Jetson TX2 graphics processing unit.

(0 min) Recent face reenactment works are limited by the coarse reference landmarks, leading to unsatisfactory identity preserving performance due to the distribution gap between the manipulated landmarks and those sampled from a real person. To address this issue, we propose a fine-grained identity-preserving landmark-guided face reenactment approach. The proposed method has two novelties. First, a landmark synthesis network which is designed to generate fine-grained landmark faces with more details. The network refines the manipulated landmarks and generates a smooth and gradually changing face landmark sequence with good identity preserving ability. Second, several novel loss functions including synthesized face identity preserving loss, foreground/background mask loss as well as boundary loss are designed, which aims at synthesizing clear and sharp high-quality faces. Experiments are conducted on our self-collected BeautySelfie and the public VoxCeleb1 datasets. The presented qualitative and quantitative results show that our method can reenact fine-grained higher quality faces with good ID-preserved appearance details, fewer artifacts and clearer boundaries than state-of-the-art works. Code will be released for reproduction.

(0 min) We consider the task of visual indoor exploration with multiple agents, where the agents need to cooperatively explore the entire indoor region using as few steps as possible. Classical planning-based methods often suffer from particularly expensive computation at each inference step and a limited expressiveness of cooperation strategy. By contrast, reinforcement learning (RL) has become a trending paradigm for tackling this challenge due to its modeling capability of arbitrarily complex strategies and minimal inference overhead. We extend the state-of-the-art single-agent RL solution, Active Neural SLAM (ANS), to the multi-agent setting by introducing a novel RL-based global-goal planner, Spatial Coordination Planner (SCP), which leverages spatial information from each individual agent in an end-to-end manner and effectively guides the agents to navigate towards different spatial goals with high exploration efficiency. SCP consists of a transformer-based relation encoder to capture intra-agent interactions and a spatial action decoder to produce accurate goals. In addition, we also implement a few multi-agent enhancements to process local information from each agent for an aligned spatial representation and more precise planning. Our final solution, Multi-Agent Active Neural SLAM (MAANS), combines all these techniques and substantially outperforms 4 different planning-based methods and various RL baselines in the photo-realistic physical testbed, Habitat.

(0 min) Test-time augmentation -- the aggregation of predictions across transformed versions of a test input -- is a common practice in image classification. Traditionally, predictions are combined using a simple average. In this paper, we present 1) experimental analyses that shed light on cases in which the simple average is suboptimal and 2) a method to address these shortcomings. A key finding is that even when test-time augmentation produces a net improvement in accuracy, it can change many correct predictions into incorrect predictions. We delve into when and why test-time augmentation changes a prediction from being correct to incorrect and vice versa. Building on these insights, we present a learning-based method for aggregating test-time augmentations. Experiments across a diverse set of models, datasets, and augmentations show that our method delivers consistent improvements over existing approaches.

(0 min) When seeing a new object, humans can immediately recognize it across different retinal locations: the internal object representation is invariant to translation. It is commonly believed that Convolutional Neural Networks (CNNs) are architecturally invariant to translation thanks to the convolution and/or pooling operations they are endowed with. In fact, several studies have found that these networks systematically fail to recognise new objects on untrained locations. In this work, we test a wide variety of CNNs architectures showing how, apart from DenseNet-121, none of the models tested was architecturally invariant to translation. Nevertheless, all of them could learn to be invariant to translation. We show how this can be achieved by pretraining on ImageNet, and it is sometimes possible with much simpler data sets when all the items are fully translated across the input canvas. At the same time, this invariance can be disrupted by further training due to catastrophic forgetting/interference. These experiments show how pretraining a network on an environment with the right `latent' characteristics (a more naturalistic environment) can result in the network learning deep perceptual rules which would dramatically improve subsequent generalization.

(0 min) In the last decade, deep learning (DL) approaches have been used successfully in computer vision (CV) applications. However, DL-based CV models are generally considered to be black boxes due to their lack of interpretability. This black box behavior has exacerbated user distrust and therefore has prevented widespread deployment DLCV models in autonomous driving tasks even though some of these models exhibit superiority over human performance. For this reason, it is essential to develop explainable DL models for autonomous driving task. Explainable DL models can not only boost user trust in autonomy but also serve as a diagnostic approach to identify anydefects and weaknesses of the model during the system development phase. In this paper, we propose an explainable end-to-end autonomous driving system based on "Transformer", a state-of-the-art (SOTA) self-attention based model, to map visual features from images collected by onboard cameras to guide potential driving actions with corresponding explanations. The model achieves a soft attention over the global features of the image. The results demonstrate the efficacy of our proposed model as it exhibits superior performance (in terms of correct prediction of actions and explanations) compared to the benchmark model by a significant margin with lower computational cost.

(0 min) Point cloud segmentation is the foundation of 3D environmental perception for modern intelligent systems. To solve this problem and image segmentation, conditional random fields (CRFs) are usually formulated as discrete models in label space to encourage label consistency, which is actually a kind of postprocessing. In this paper, we reconsider the CRF in feature space for point cloud segmentation because it can capture the structure of features well to improve the representation ability of features rather than simply smoothing. Therefore, we first model the point cloud features with a continuous quadratic energy model and formulate its solution process as a message-passing graph convolution, by which it can be easily integrated into a deep network. We theoretically demonstrate that the message passing in the graph convolution is equivalent to the mean-field approximation of a continuous CRF model. Furthermore, we build an encoder-decoder network based on the proposed continuous CRF graph convolution (CRFConv), in which the CRFConv embedded in the decoding layers can restore the details of high-level features that were lost in the encoding stage to enhance the location ability of the network, thereby benefiting segmentation. Analogous to the CRFConv, we show that the classical discrete CRF can also work collaboratively with the proposed network via another graph convolution to further improve the segmentation results. Experiments on various point cloud benchmarks demonstrate the effectiveness and robustness of the proposed method. Compared with the state-of-the-art methods, the proposed method can also achieve competitive segmentation performance.

(0 min) Growing at a fast pace, modern autonomous systems will soon be deployed at scale, opening up the possibility for cooperative multi-agent systems. Sharing information and distributing workloads allow autonomous agents to better perform tasks and increase computation efficiency. However, shared information can be modified to execute adversarial attacks on deep learning models that are widely employed in modern systems. Thus, we aim to study the robustness of such systems and focus on exploring adversarial attacks in a novel multi-agent setting where communication is done through sharing learned intermediate representations of neural networks. We observe that an indistinguishable adversarial message can severely degrade performance, but becomes weaker as the number of benign agents increases. Furthermore, we show that black-box transfer attacks are more difficult in this setting when compared to directly perturbing the inputs, as it is necessary to align the distribution of learned representations with domain adaptation. Our work studies robustness at the neural network level to contribute an additional layer of fault tolerance to modern security protocols for more secure multi-agent systems.

(0 min) Mitral valve repair is a very difficult operation, often requiring experienced surgeons. The doctor will insert a prosthetic ring to aid in the restoration of heart function. The location of the prosthesis' sutures is critical. Obtaining and studying them during the procedure is a valuable learning experience for new surgeons. This paper proposes a landmark detection network for detecting sutures in endoscopic pictures, which solves the problem of a variable number of suture points in the images. Because there are two datasets, one from the simulated domain and the other from real intraoperative data, this work uses cycleGAN to interconvert the images from the two domains to obtain a larger dataset and a better score on real intraoperative data. This paper performed the tests using a simulated dataset of 2708 photos and a real dataset of 2376 images. The mean sensitivity on the simulated dataset is about 75.64% and the precision is about 73.62%. The mean sensitivity on the real dataset is about 50.23% and the precision is about 62.76%. The data is from the AdaptOR MICCAI Challenge 2021, which can be found at https://zenodo.org/record/4646979\#.YO1zLUxCQ2x.

(0 min) Image deblurring is a classical computer vision problem that aims to recover a sharp image from a blurred image. To solve this problem, existing methods apply the Encode-Decode architecture to design the complex networks to make a good performance. However, most of these methods use repeated up-sampling and down-sampling structures to expand the receptive field, which results in texture information loss during the sampling process and some of them design the multiple stages that lead to difficulties with convergence. Therefore, our model uses dilated convolution to enable the obtainment of the large receptive field with high spatial resolution. Through making full use of the different receptive fields, our method can achieve better performance. On this basis, we reduce the number of up-sampling and down-sampling and design a simple network structure. Besides, we propose a novel module using the wavelet transform, which effectively helps the network to recover clear high-frequency texture details. Qualitative and quantitative evaluations of real and synthetic datasets show that our deblurring method is comparable to existing algorithms in terms of performance with much lower training requirements. The source code and pre-trained models are available at https://github.com/FlyEgle/SDWNet.

(0 min) The ability to predict future outcomes conditioned on observed video frames is crucial for intelligent decision-making in autonomous systems. Recently, deep recurrent architectures have been applied to the task of video prediction. However, this often results in blurry predictions and requires tedious training on large datasets. Here, we explore a different approach by (1) using frequency-domain approaches for video prediction and (2) explicitly inferring object-motion relationships in the observed scene. The resulting predictions are consistent with the observed dynamics in a scene and do not suffer from blur.

(0 min) 3D shape reconstruction is essential in the navigation of minimally-invasive and auto robot-guided surgeries whose operating environments are indirect and narrow, and there have been some works that focused on reconstructing the 3D shape of the surgical organ through limited 2D information available. However, the lack and incompleteness of such information caused by intraoperative emergencies (such as bleeding) and risk control conditions have not been considered. In this paper, a novel hierarchical shape-perception network (HSPN) is proposed to reconstruct the 3D point clouds (PCs) of specific brains from one single incomplete image with low latency. A branching predictor and several hierarchical attention pipelines are constructed to generate point clouds that accurately describe the incomplete images and then complete these point clouds with high quality. Meanwhile, attention gate blocks (AGBs) are designed to efficiently aggregate geometric local features of incomplete PCs transmitted by hierarchical attention pipelines and internal features of reconstructing point clouds. With the proposed HSPN, 3D shape perception and completion can be achieved spontaneously. Comprehensive results measured by Chamfer distance and PC-to-PC error demonstrate that the performance of the proposed HSPN outperforms other competitive methods in terms of qualitative displays, quantitative experiment, and classification evaluation.

(0 min) Despite significant progress, we show that state of the art 3D human pose and shape estimation methods remain sensitive to partial occlusion and can produce dramatically wrong predictions although much of the body is observable. To address this, we introduce a soft attention mechanism, called the Part Attention REgressor (PARE), that learns to predict body-part-guided attention masks. We observe that state-of-the-art methods rely on global feature representations, making them sensitive to even small occlusions. In contrast, PARE's part-guided attention mechanism overcomes these issues by exploiting information about the visibility of individual body parts while leveraging information from neighboring body-parts to predict occluded parts. We show qualitatively that PARE learns sensible attention masks, and quantitative evaluation confirms that PARE achieves more accurate and robust reconstruction results than existing approaches on both occlusion-specific and standard benchmarks. The code and data are available for research purposes at {\small \url{https://pare.is.tue.mpg.de/}}

(0 min) Stereo image pairs encode 3D scene cues into stereo correspondences between the left and right images. To exploit 3D cues within stereo images, recent CNN based methods commonly use cost volume techniques to capture stereo correspondence over large disparities. However, since disparities can vary significantly for stereo cameras with different baselines, focal lengths and resolutions, the fixed maximum disparity used in cost volume techniques hinders them to handle different stereo image pairs with large disparity variations. In this paper, we propose a generic parallax-attention mechanism (PAM) to capture stereo correspondence regardless of disparity variations. Our PAM integrates epipolar constraints with attention mechanism to calculate feature similarities along the epipolar line to capture stereo correspondence. Based on our PAM, we propose a parallax-attention stereo matching network (PASMnet) and a parallax-attention stereo image super-resolution network (PASSRnet) for stereo matching and stereo image super-resolution tasks. Moreover, we introduce a new and large-scale dataset named Flickr1024 for stereo image super-resolution. Experimental results show that our PAM is generic and can effectively learn stereo correspondence under large disparity variations in an unsupervised manner. Comparative results show that our PASMnet and PASSRnet achieve the state-of-the-art performance.

(0 min) Spatial convolutions are widely used in numerous deep video models. It fundamentally assumes spatio-temporal invariance, i.e., using shared weights for every location in different frames. This work presents Temporally-Adaptive Convolutions (TAdaConv) for video understanding, which shows that adaptive weight calibration along the temporal dimension is an efficient way to facilitate modelling complex temporal dynamics in videos. Specifically, TAdaConv empowers the spatial convolutions with temporal modelling abilities by calibrating the convolution weights for each frame according to its local and global temporal context. Compared to previous temporal modelling operations, TAdaConv is more efficient as it operates over the convolution kernels instead of the features, whose dimension is an order of magnitude smaller than the spatial resolutions. Further, the kernel calibration also brings an increased model capacity. We construct TAda2D networks by replacing the spatial convolutions in ResNet with TAdaConv, which leads to on par or better performance compared to state-of-the-art approaches on multiple video action recognition and localization benchmarks. We also demonstrate that as a readily plug-in operation with negligible computation overhead, TAdaConv can effectively improve many existing video models with a convincing margin. Codes and models will be made available at https://github.com/alibaba-mmai-research/pytorch-video-understanding.

(0 min) In contrast to Connectionist Temporal Classification (CTC) approaches, Sequence-To-Sequence (S2S) models for Handwritten Text Recognition (HTR) suffer from errors such as skipped or repeated words which often occur at the end of a sequence. In this paper, to combine the best of both approaches, we propose to use the CTC-Prefix-Score during S2S decoding. Hereby, during beam search, paths that are invalid according to the CTC confidence matrix are penalised. Our network architecture is composed of a Convolutional Neural Network (CNN) as visual backbone, bidirectional Long-Short-Term-Memory-Cells (LSTMs) as encoder, and a decoder which is a Transformer with inserted mutual attention layers. The CTC confidences are computed on the encoder while the Transformer is only used for character-wise S2S decoding. We evaluate this setup on three HTR data sets: IAM, Rimes, and StAZH. On IAM, we achieve a competitive Character Error Rate (CER) of 2.95% when pretraining our model on synthetic data and including a character-based language model for contemporary English. Compared to other state-of-the-art approaches, our model requires about 10-20 times less parameters. Access our shared implementations via this link to GitHub: https://github.com/Planet-AI-GmbH/tfaip-hybrid-ctc-s2s.

(0 min) We introduce Amazon-Berkeley Objects (ABO), a new large-scale dataset of product images and 3D models corresponding to real household objects. We use this realistic, object-centric 3D dataset to measure the domain gap for single-view 3D reconstruction networks trained on synthetic objects. We also use multi-view images from ABO to measure the robustness of state-of-the-art metric learning approaches to different camera viewpoints. Finally, leveraging the physically-based rendering materials in ABO, we perform single- and multi-view material estimation for a variety of complex, real-world geometries. The full dataset is available for download at https://amazon-berkeley-objects.s3.amazonaws.com/index.html.

(0 min) In the field of action recognition, video clips are always treated as ordered frames for subsequent processing. To achieve spatio-temporal perception, existing approaches propose to embed adjacent temporal interaction in the convolutional layer. The global semantic information can therefore be obtained by stacking multiple local layers hierarchically. However, such global temporal accumulation can only reflect the high-level semantics in deep layers, neglecting the potential low-level holistic clues in shallow layers. In this paper, we first propose to transform a video sequence into a graph to obtain direct long-term dependencies among temporal frames. To preserve sequential information during transformation, we devise a structured graph module (SGM), achieving fine-grained temporal interactions throughout the entire network. In particular, SGM divides the neighbors of each node into several temporal regions so as to extract global structural information with diverse sequential flows. Extensive experiments are performed on standard benchmark datasets, i.e., Something-Something V1 & V2, Diving48, Kinetics-400, UCF101, and HMDB51. The reported performance and analysis demonstrate that SGM can achieve outstanding precision with less computational complexity.

(0 min) Faster inference of deep learning models is highly demanded on edge devices and even servers, for both financial and environmental reasons. To address this issue, we propose SoftNeuro, a novel, high-performance inference framework with efficient performance tuning. The key idea is to separate algorithmic routines from network layers. Our framework maximizes the inference performance by profiling various routines for each layer and selecting the fastest path. To efficiently find the best path, we propose a routine-selection algorithm based on dynamic programming. Experiments show that the proposed framework achieves both fast inference and efficient tuning.

(0 min) It is crucial to understand the robustness of text detection models with regard to extensive corruptions, since scene text detection techniques have many practical applications. For systematically exploring this problem, we propose two datasets from which to evaluate scene text detection models: ICDAR2015-C (IC15-C) and CTW1500-C (CTW-C). Our study extends the investigation of the performance and robustness of the proposed region proposal, regression and segmentation-based scene text detection frameworks. Furthermore, we perform a robustness analysis of six key components: pre-training data, backbone, feature fusion module, multi-scale predictions, representation of text instances and loss function. Finally, we present a simple yet effective data-based method to destroy the smoothness of text regions by merging background and foreground, which can significantly increase the robustness of different text detection networks. We hope that this study will provide valid data points as well as experience for future research. Benchmark, code and data will be made available at \url{https://github.com/wushilian/robust-scene-text-detection-benchmark}.

(0 min) AI technologies for Natural Languages have made tremendous progress recently. However, commensurate progress has not been made on Sign Languages, in particular, in recognizing signs as individual words or as complete sentences. We introduce OpenHands, a library where we take four key ideas from the NLP community for low-resource languages and apply them to sign languages for word-level recognition. First, we propose using pose extracted through pretrained models as the standard modality of data to reduce training time and enable efficient inference, and we release standardized pose datasets for 6 different sign languages - American, Argentinian, Chinese, Greek, Indian, and Turkish. Second, we train and release checkpoints of 4 pose-based isolated sign language recognition models across all 6 languages, providing baselines and ready checkpoints for deployment. Third, to address the lack of labelled data, we propose self-supervised pretraining on unlabelled data. We curate and release the largest pose-based pretraining dataset on Indian Sign Language (Indian-SL). Fourth, we compare different pretraining strategies and for the first time establish that pretraining is effective for sign language recognition by demonstrating (a) improved fine-tuning performance especially in low-resource settings, and (b) high crosslingual transfer from Indian-SL to few other sign languages. We open-source all models and datasets in OpenHands with a hope that it makes research in sign languages more accessible, available here at https://github.com/AI4Bharat/OpenHands .

(0 min) Monocular depth estimation is the base task in computer vision. It has a tremendous development in the decade with the development of deep learning. But the boundary blur of the depth map is still a serious problem. Research finds the boundary blur problem is mainly caused by two factors, first, the low-level features containing boundary and structure information may loss in deeper networks during the convolution process., second, the model ignores the errors introduced by the boundary area due to the few portions of the boundary in the whole areas during the backpropagation. In order to mitigate the boundary blur problem, we focus on the above two impact factors. Firstly, we design a scene understanding module to learn the global information with low- and high-level features, and then to transform the global information to different scales with our proposed scale transform module according to the different phases in the decoder. Secondly, we propose a boundary-aware depth loss function to pay attention to the effects of the boundary's depth value. The extensive experiments show that our method can predict the depth maps with clearer boundaries, and the performance of the depth accuracy base on NYU-depth v2 and SUN RGB-D is competitive.

(0 min) Scene graph is a structured representation of a scene that can clearly express the objects, attributes, and relationships between objects in the scene. As computer vision technology continues to develop, people are no longer satisfied with simply detecting and recognizing objects in images; instead, people look forward to a higher level of understanding and reasoning about visual scenes. For example, given an image, we want to not only detect and recognize objects in the image, but also know the relationship between objects (visual relationship detection), and generate a text description (image captioning) based on the image content. Alternatively, we might want the machine to tell us what the little girl in the image is doing (Visual Question Answering (VQA)), or even remove the dog from the image and find similar images (image editing and retrieval), etc. These tasks require a higher level of understanding and reasoning for image vision tasks. The scene graph is just such a powerful tool for scene understanding. Therefore, scene graphs have attracted the attention of a large number of researchers, and related research is often cross-modal, complex, and rapidly developing. However, no relatively systematic survey of scene graphs exists at present. To this end, this survey conducts a comprehensive investigation of the current scene graph research. More specifically, we first summarized the general definition of the scene graph, then conducted a comprehensive and systematic discussion on the generation method of the scene graph (SGG) and the SGG with the aid of prior knowledge. We then investigated the main applications of scene graphs and summarized the most commonly used datasets. Finally, we provide some insights into the future development of scene graphs. We believe this will be a very helpful foundation for future research on scene graphs.

(0 min) In this paper we exploit Semi-Supervised Learning (SSL) to increase the amount of training data to improve the performance of Fine-Grained Visual Categorization (FGVC). This problem has not been investigated in the past in spite of prohibitive annotation costs that FGVC requires. Our approach leverages unlabeled data with an adversarial optimization strategy in which the internal features representation is obtained with a second-order pooling model. This combination allows to back-propagate the information of the parts, represented by second-order pooling, onto unlabeled data in an adversarial training setting. We demonstrate the effectiveness of the combined use by conducting experiments on six state-of-the-art fine-grained datasets, which include Aircrafts, Stanford Cars, CUB-200-2011, Oxford Flowers, Stanford Dogs, and the recent Semi-Supervised iNaturalist-Aves. Experimental results clearly show that our proposed method has better performance than the only previous approach that examined this problem; it also obtained higher classification accuracy with respect to the supervised learning methods with which we compared.

(0 min) Binary Neural Networks (BNNs) rely on a real-valued auxiliary variable W to help binary training. However, pioneering binary works only use W to accumulate gradient updates during backward propagation, which can not fully exploit its power and may hinder novel advances in BNNs. In this work, we explore the role of W in training besides acting as a latent variable. Notably, we propose to add W into the computation graph, making it perform as a real-valued feature extractor to aid the binary training. We make different attempts on how to utilize the real-valued weights and propose a specialized supervision. Visualization experiments qualitatively verify the effectiveness of our approach in making it easier to distinguish between different categories. Quantitative experiments show that our approach outperforms current state-of-the-arts, further closing the performance gap between floating-point networks and BNNs. Evaluation on ImageNet with ResNet-18 (Top-1 63.4%), ResNet-34 (Top-1 67.0%) achieves new state-of-the-art.

(0 min) In this paper, we propose a method for the automatic semantic segmentation of satellite images into six classes (sparse forest, dense forest, moor, herbaceous formation, building, and road). We rely on Swin Transformer architecture and build the dataset from IGN open data. We report quantitative and qualitative segmentation results on this dataset and discuss strengths and limitations. The dataset and the trained model are made publicly available.

(0 min) Unsupervised learning of depth from indoor monocular videos is challenging as the artificial environment contains many textureless regions. Fortunately, the indoor scenes are full of specific structures, such as planes and lines, which should help guide unsupervised depth learning. This paper proposes PLNet that leverages the plane and line priors to enhance the depth estimation. We first represent the scene geometry using local planar coefficients and impose the smoothness constraint on the representation. Moreover, we enforce the planar and linear consistency by randomly selecting some sets of points that are probably coplanar or collinear to construct simple and effective consistency losses. To verify the proposed method's effectiveness, we further propose to evaluate the flatness and straightness of the predicted point cloud on the reliable planar and linear regions. The regularity of these regions indicates quality indoor reconstruction. Experiments on NYU Depth V2 and ScanNet show that PLNet outperforms existing methods. The code is available at \url{https://github.com/HalleyJiang/PLNet}.

(0 min) The pretrain-finetune paradigm has shown outstanding performance on many applications of deep learning, where a model is pre-trained on a upstream large dataset (e.g. ImageNet), and is then fine-tuned to different downstream tasks. Though for most cases, the pre-training stage is conducted based on supervised methods, recent works on self-supervised pre-training have shown powerful transferability and even outperform supervised pre-training on multiple downstream tasks. It thus remains an open question how to better generalize supervised pre-training model to downstream tasks. In this paper, we argue that the worse transferability of existing supervised pre-training methods arise from the negligence of valuable intra-class semantic difference. This is because these methods tend to push images from the same class close to each other despite of the large diversity in their visual contents, a problem to which referred as "overfit of upstream tasks". To alleviate this problem, we propose a new supervised pre-training method based on Leave-One-Out K-Nearest-Neighbor, or LOOK for short. It relieves the problem of overfitting upstream tasks by only requiring each image to share its class label with most of its k nearest neighbors, thus allowing each class to exhibit a multi-mode distribution and consequentially preserving part of intra-class difference for better transferring to downstream tasks. We developed efficient implementation of the proposed method that scales well to large datasets. Experimental studies on multiple downstream tasks show that LOOK outperforms other state-of-the-art methods for supervised and self-supervised pre-training.

(0 min) We present a two-stage learning framework for weakly supervised object localization (WSOL). While most previous efforts rely on high-level feature based CAMs (Class Activation Maps), this paper proposes to localize objects using the low-level feature based activation maps. In the first stage, an activation map generator produces activation maps based on the low-level feature maps in the classifier, such that rich contextual object information is included in an online manner. In the second stage, we employ an evaluator to evaluate the activation maps predicted by the activation map generator. Based on this, we further propose a weighted entropy loss, an attentive erasing, and an area loss to drive the activation map generator to substantially reduce the uncertainty of activations between object and background, and explore less discriminative regions. Based on the low-level object information preserved in the first stage, the second stage model gradually generates a well-separated, complete, and compact activation map of object in the image, which can be easily thresholded for accurate localization. Extensive experiments on CUB-200-2011 and ImageNet-1K datasets show that our framework surpasses previous methods by a large margin, which sets a new state-of-the-art for WSOL.

(0 min) Dermatologists often diagnose or rule out early melanoma by evaluating the follow-up dermoscopic images of skin lesions. However, existing algorithms for early melanoma diagnosis are developed using single time-point images of lesions. Ignoring the temporal, morphological changes of lesions can lead to misdiagnosis in borderline cases. In this study, we propose a framework for automated early melanoma diagnosis using sequential dermoscopic images. To this end, we construct our method in three steps. First, we align sequential dermoscopic images of skin lesions using estimated Euclidean transformations, extract the lesion growth region by computing image differences among the consecutive images, and then propose a spatio-temporal network to capture the dermoscopic changes from aligned lesion images and the corresponding difference images. Finally, we develop an early diagnosis module to compute probability scores of malignancy for lesion images over time. We collected 179 serial dermoscopic imaging data from 122 patients to verify our method. Extensive experiments show that the proposed model outperforms other commonly used sequence models. We also compared the diagnostic results of our model with those of seven experienced dermatologists and five registrars. Our model achieved higher diagnostic accuracy than clinicians (63.69% vs. 54.33%, respectively) and provided an earlier diagnosis of melanoma (60.7% vs. 32.7% of melanoma correctly diagnosed on the first follow-up images). These results demonstrate that our model can be used to identify melanocytic lesions that are at high-risk of malignant transformation earlier in the disease process and thereby redefine what is possible in the early detection of melanoma.

(0 min) Human epidermal growth factor receptor 2 (HER2) is an important prognostic and predictive factor which is overexpressed in 15-20% of breast cancer (BCa). The determination of its status is a key clinical decision making step for selection of treatment regimen and prognostication. HER2 status is evaluated using transcroptomics or immunohistochemistry (IHC) through situ hybridisation (ISH) which require additional costs and tissue burden in addition to analytical variabilities in terms of manual observational biases in scoring. In this study, we propose a novel graph neural network (GNN) based model (termed SlideGraph+) to predict HER2 status directly from whole-slide images of routine Haematoxylin and Eosin (H&E) slides. The network was trained and tested on slides from The Cancer Genome Atlas (TCGA) in addition to two independent test datasets. We demonstrate that the proposed model outperforms the state-of-the-art methods with area under the ROC curve (AUC) values > 0.75 on TCGA and 0.8 on independent test sets. Our experiments show that the proposed approach can be utilised for case triaging as well as pre-ordering diagnostic tests in a diagnostic setting. It can also be used for other weakly supervised prediction problems in computational pathology. The SlideGraph+ code is available at https://github.com/wenqi006/SlideGraph.

(0 min) The importance and relevance of digital image forensics has attracted researchers to establish different techniques for creating as well as detecting forgeries. The core category in passive image forgery is copy-move image forgery that affects the originality of image by applying a different transformation. In this paper frequency domain image manipulation method is being presented.The method exploits the localized nature of discrete wavelet transform (DWT) to get hold of the region of the host image to be manipulated. Both the patch and host image are subjected to DWT at the same level $l$ to get $3l + 1$ sub-bands and each sub-band of the patch is pasted to the identified region in the corresponding sub-band of the host image. The resultant manipulated host sub-bands are then subjected to inverse DWT to get the final manipulated host image. The proposed method shows good resistance against detection by two frequency domain forgery detection methods from the literature. The purpose of this research work is to create the forgery and highlight the need to produce forgery detection methods that are robust against the malicious copy-move forgery.

(0 min) As a challenging task, unsupervised person ReID aims to match the same identity with query images which does not require any labeled information. In general, most existing approaches focus on the visual cues only, leaving potentially valuable auxiliary metadata information (e.g., spatio-temporal context) unexplored. In the real world, such metadata is normally available alongside captured images, and thus plays an important role in separating several hard ReID matches. With this motivation in mind, we propose~\textbf{MGH}, a novel unsupervised person ReID approach that uses meta information to construct a hypergraph for feature learning and label refinement. In principle, the hypergraph is composed of camera-topology-aware hyperedges, which can model the heterogeneous data correlations across cameras. Taking advantage of label propagation on the hypergraph, the proposed approach is able to effectively refine the ReID results, such as correcting the wrong labels or smoothing the noisy labels. Given the refined results, We further present a memory-based listwise loss to directly optimize the average precision in an approximate manner. Extensive experiments on three benchmarks demonstrate the effectiveness of the proposed approach against the state-of-the-art.

(0 min) "The power of a generalization system follows directly from its biases" (Mitchell 1980). Today, CNNs are incredibly powerful generalisation systems -- but to what degree have we understood how their inductive bias influences model decisions? We here attempt to disentangle the various aspects that determine how a model decides. In particular, we ask: what makes one model decide differently from another? In a meticulously controlled setting, we find that (1.) irrespective of the network architecture or objective (e.g. self-supervised, semi-supervised, vision transformers, recurrent models) all models end up with a similar decision boundary. (2.) To understand these findings, we analysed model decisions on the ImageNet validation set from epoch to epoch and image by image. We find that the ImageNet validation set, among others, suffers from dichotomous data difficulty (DDD): For the range of investigated models and their accuracies, it is dominated by 46.0% "trivial" and 11.5% "impossible" images (beyond label errors). Only 42.5% of the images could possibly be responsible for the differences between two models' decision boundaries. (3.) Only removing the "impossible" and "trivial" images allows us to see pronounced differences between models. (4.) Humans are highly accurate at predicting which images are "trivial" and "impossible" for CNNs (81.4%). This implies that in future comparisons of brains, machines and behaviour, much may be gained from investigating the decisive role of images and the distribution of their difficulties.

(0 min) Neural Architecture Search (NAS) represents an emerging machine learning (ML) paradigm that automatically searches for models tailored to given tasks, which greatly simplifies the development of ML systems and propels the trend of ML democratization. Yet, little is known about the potential security risks incurred by NAS, which is concerning given the increasing use of NAS-generated models in critical domains. This work represents a solid initial step towards bridging the gap. Through an extensive empirical study of 10 popular NAS methods, we show that compared with their manually designed counterparts, NAS-generated models tend to suffer greater vulnerability to various malicious attacks (e.g., adversarial evasion, model poisoning, and functionality stealing). Further, with both empirical and analytical evidence, we provide possible explanations for such phenomena: given the prohibitive search space and training cost, most NAS methods favor models that converge fast at early training stages; this preference results in architectural properties associated with attack vulnerability (e.g., high loss smoothness and low gradient variance). Our findings not only reveal the relationships between model characteristics and attack vulnerability but also suggest the inherent connections underlying different attacks. Finally, we discuss potential remedies to mitigate such drawbacks, including increasing cell depth and suppressing skip connects, which lead to several promising research directions.

(0 min) After significant earthquakes, we can see images posted on social media platforms by individuals and media agencies owing to the mass usage of smartphones these days. These images can be utilized to provide information about the shaking damage in the earthquake region both to the public and research community, and potentially to guide rescue work. This paper presents an automated way to extract the damaged building images after earthquakes from social media platforms such as Twitter and thus identify the particular user posts containing such images. Using transfer learning and ~6500 manually labelled images, we trained a deep learning model to recognize images with damaged buildings in the scene. The trained model achieved good performance when tested on newly acquired images of earthquakes at different locations and ran in near real-time on Twitter feed after the 2020 M7.0 earthquake in Turkey. Furthermore, to better understand how the model makes decisions, we also implemented the Grad-CAM method to visualize the important locations on the images that facilitate the decision.

(0 min) This paper sustains the position that the time has come for thinking of learning machines that conquer visual skills in a truly human-like context, where a few human-like object supervisions are given by vocal interactions and pointing aids only. This likely requires new foundations on computational processes of vision with the final purpose of involving machines in tasks of visual description by living in their own visual environment under simple man-machine linguistic interactions. The challenge consists of developing machines that learn to see without needing to handle visual databases. This might open the doors to a truly orthogonal competitive track concerning deep learning technologies for vision which does not rely on the accumulation of huge visual databases.

(0 min) Correspondence-based point cloud registration is a cornerstone in geometric computer vision, robotics perception, photogrammetry and remote sensing, which seeks to estimate the best rigid transformation between two point clouds from the correspondences established over 3D keypoints. However, due to limited robustness and accuracy, current 3D keypoint matching techniques are very prone to yield outliers, probably even in very large numbers, making robust estimation for point cloud registration of great importance. Unfortunately, existing robust methods may suffer from high computational cost or insufficient robustness when encountering high (or even extreme) outlier ratios, hardly ideal enough for practical use. In this paper, we present a novel time-efficient RANSAC-type consensus maximization solver, named DANIEL (Double-layered sAmpliNg with consensus maximization based on stratIfied Element-wise compatibiLity), for robust registration. DANIEL is designed with two layers of random sampling, in order to find inlier subsets with the lowest computational cost possible. Specifically, we: (i) apply the rigidity constraint to prune raw outliers in the first layer of one-point sampling, (ii) introduce a series of stratified element-wise compatibility tests to conduct rapid compatibility checking between minimal models so as to realize more efficient consensus maximization in the second layer of two-point sampling, and (iii) probabilistic termination conditions are employed to ensure the timely return of the final inlier set. Based on a variety of experiments over multiple real datasets, we show that DANIEL is robust against over 99% outliers and also significantly faster than existing state-of-the-art robust solvers (e.g. RANSAC, FGR, GORE).

(0 min) Vision-and-language navigation (VLN) is a trending topic which aims to navigate an intelligent agent to an expected position through natural language instructions. This work addresses the task of VLN from a previously-ignored aspect, namely the spatial route prior of the navigation scenes. A critically enabling innovation of this work is explicitly considering the spatial route prior under several different VLN settings. In a most information-rich case of knowing environment maps and admitting shortest-path prior, we observe that given an origin-destination node pair, the internal route can be uniquely determined. Thus, VLN can be effectively formulated as an ordinary classification problem over all possible destination nodes in the scenes. Furthermore, we relax it to other more general VLN settings, proposing a sequential-decision variant (by abandoning the shortest-path route prior) and an explore-and-exploit scheme (for addressing the case of not knowing the environment maps) that curates a compact and informative sub-graph to exploit. As reported by [34], the performance of VLN methods has been stuck at a plateau in past two years. Even with increased model complexity, the state-of-the-art success rate on R2R validation-unseen set has stayed around 62% for single-run and 73% for beam-search with model-ensemble. We have conducted comprehensive evaluations on both R2R and R4R, and surprisingly found that utilizing the spatial route priors may be the key of breaking above-mentioned performance ceiling. For example, on R2R validation-unseen set, when the number of discrete nodes explored is about 40, our single-model success rate reaches 73%, and increases to 78% if a Speaker model is ensembled, which significantly outstrips previous state-of-the-art VLN-BERT with 3 models ensembled.

(0 min) Retrieving accurate 3D reconstructions of objects from the way they reflect light is a very challenging task in computer vision. Despite more than four decades since the definition of the Photometric Stereo problem, most of the literature has had limited success when global illumination effects such as cast shadows, self-reflections and ambient light come into play, especially for specular surfaces. Recent approaches have leveraged the power of deep learning in conjunction with computer graphics in order to cope with the need of a vast number of training data in order to invert the image irradiance equation and retrieve the geometry of the object. However, rendering global illumination effects is a slow process which can limit the amount of training data that can be generated. In this work we propose a novel pixel-wise training procedure for normal prediction by replacing the training data (observation maps) of globally rendered images with independent per-pixel generated data. We show that global physical effects can be approximated on the observation map domain and this simplifies and speeds up the data creation procedure. Our network, PX-NET, achieves the state-of-the-art performance compared to other pixelwise methods on synthetic datasets, as well as the Diligent real dataset on both dense and sparse light settings.

(0 min) We observe that despite their hierarchical convolutional nature, the synthesis process of typical generative adversarial networks depends on absolute pixel coordinates in an unhealthy manner. This manifests itself as, e.g., detail appearing to be glued to image coordinates instead of the surfaces of depicted objects. We trace the root cause to careless signal processing that causes aliasing in the generator network. Interpreting all signals in the network as continuous, we derive generally applicable, small architectural changes that guarantee that unwanted information cannot leak into the hierarchical synthesis process. The resulting networks match the FID of StyleGAN2 but differ dramatically in their internal representations, and they are fully equivariant to translation and rotation even at subpixel scales. Our results pave the way for generative models better suited for video and animation.

(0 min) Data augmentation has been an indispensable tool to improve the performance of deep neural networks, however the augmentation can hardly transfer among different tasks and datasets. Consequently, a recent trend is to adopt AutoML technique to learn proper augmentation policy without extensive hand-crafted tuning. In this paper, we propose an efficient differentiable search algorithm called Direct Differentiable Augmentation Search (DDAS). It exploits meta-learning with one-step gradient update and continuous relaxation to the expected training loss for efficient search. Our DDAS can achieve efficient augmentation search without relying on approximations such as Gumbel Softmax or second order gradient approximation. To further reduce the adverse effect of improper augmentations, we organize the search space into a two level hierarchy, in which we first decide whether to apply augmentation, and then determine the specific augmentation policy. On standard image classification benchmarks, our DDAS achieves state-of-the-art performance and efficiency tradeoff while reducing the search cost dramatically, e.g. 0.15 GPU hours for CIFAR-10. In addition, we also use DDAS to search augmentation for object detection task and achieve comparable performance with AutoAugment, while being 1000x faster.

(0 min) Recently introduced implicit field representations offer an effective way of generating 3D object shapes. They leverage implicit decoder trained to take a 3D point coordinate concatenated with a shape encoding and to output a value which indicates whether the point is outside the shape or not. Although this approach enables efficient rendering of visually plausible objects, it has two significant limitations. First, it is based on a single neural network dedicated for all objects from a training set which results in a cumbersome training procedure and its application in real life. More importantly, the implicit decoder takes only points sampled within voxels (and not the entire voxels) which yields problems at the classification boundaries and results in empty spaces within the rendered mesh. To solve the above limitations, we introduce a new HyperCube architecture based on interval arithmetic network, that enables direct processing of 3D voxels, trained using a hypernetwork paradigm to enforce model convergence. Instead of processing individual 3D samples from within a voxel, our approach allows to input the entire voxel (3D cube) represented with its convex hull coordinates, while the target network constructed by a hypernet assigns it to an inside or outside category. As a result our HyperCube model outperforms the competing approaches both in terms of training and inference efficiency, as well as the final mesh quality.

(0 min) Communication between embodied AI agents has received increasing attention in recent years. Despite its use, it is still unclear whether the learned communication is interpretable and grounded in perception. To study the grounding of emergent forms of communication, we first introduce the collaborative multi-object navigation task CoMON. In this task, an oracle agent has detailed environment information in the form of a map. It communicates with a navigator agent that perceives the environment visually and is tasked to find a sequence of goals. To succeed at the task, effective communication is essential. CoMON hence serves as a basis to study different communication mechanisms between heterogeneous agents, that is, agents with different capabilities and roles. We study two common communication mechanisms and analyze their communication patterns through an egocentric and spatial lens. We show that the emergent communication can be grounded to the agent observations and the spatial structure of the 3D environment. Video summary: https://youtu.be/kLv2rxO9t0g

(0 min) If an image tells a story, the image caption is the briefest narrator. Generally, a scene graph prefers to be an omniscient generalist, while the image caption is more willing to be a specialist, which outlines the gist. Lots of previous studies have found that a scene graph is not as practical as expected unless it can reduce the trivial contents and noises. In this respect, the image caption is a good tutor. To this end, we let the scene graph borrow the ability from the image caption so that it can be a specialist on the basis of remaining all-around, resulting in the so-called Topic Scene Graph. What an image caption pays attention to is distilled and passed to the scene graph for estimating the importance of partial objects, relationships, and events. Specifically, during the caption generation, the attention about individual objects in each time step is collected, pooled, and assembled to obtain the attention about relationships, which serves as weak supervision for regularizing the estimated importance scores of relationships. In addition, as this attention distillation process provides an opportunity for combining the generation of image caption and scene graph together, we further transform the scene graph into linguistic form with rich and free-form expressions by sharing a single generation model with image caption. Experiments show that attention distillation brings significant improvements in mining important relationships without strong supervision, and the topic scene graph shows great potential in subsequent applications.

(0 min) Modern deep neural networks are highly over-parameterized compared to the data on which they are trained, yet they often generalize remarkably well. A flurry of recent work has asked: why do deep networks not overfit to their training data? In this work, we make a series of empirical observations that investigate the hypothesis that deeper networks are inductively biased to find solutions with lower rank embeddings. We conjecture that this bias exists because the volume of functions that maps to low-rank embedding increases with depth. We show empirically that our claim holds true on finite width linear and non-linear models and show that these are the solutions that generalize well. We then show that the low-rank simplicity bias exists even after training, using a wide variety of commonly used optimizers. We found this phenomenon to be resilient to initialization, hyper-parameters, and learning methods. We further demonstrate how linear over-parameterization of deep non-linear models can be used to induce low-rank bias, improving generalization performance without changing the effective model capacity. Practically, we demonstrate that simply linearly over-parameterizing standard models at training time can improve performance on image classification tasks, including ImageNet.

(2 min) Previous image forensics schemes for crop detection are only limited on predicting whether an image has been cropped. This paper presents a novel scheme for image crop localization using robust watermarking. We further extend our scheme to detect tampering attack on the attacked image. We demonstrate that our scheme is the first to provide high-accuracy and robust image crop localization. Besides, the accuracy of tamper detection is comparable to many state-of-the-art methods.

(2 min) This paper focuses on task recognition and action segmentation in weakly-labeled instructional videos, where only the ordered sequence of video-level actions is available during training. We propose a two-stream framework, which exploits semantic and temporal hierarchies to recognize top-level tasks in instructional videos. Further, we present a novel top-down weakly-supervised action segmentation approach, where the predicted task is used to constrain the inference of fine-grained action sequences. Experimental results on the popular Breakfast and Cooking 2 datasets show that our two-stream hierarchical task modeling significantly outperforms existing methods in top-level task recognition for all datasets and metrics. Additionally, using our task recognition framework in the proposed top-down action segmentation approach consistently improves the state of the art, while also reducing segmentation inference time by 80-90 percent.

(2 min) We present a modern solution to the multi-view photometric stereo problem (MVPS). Our work suitably exploits the image formation model in a MVPS experimental setup to recover the dense 3D reconstruction of an object from images. We procure the surface orientation using a photometric stereo (PS) image formation model and blend it with a multi-view neural radiance field representation to recover the object's surface geometry. Contrary to the previous multi-staged framework to MVPS, where the position, iso-depth contours, or orientation measurements are estimated independently and then fused later, our method is simple to implement and realize. Our method performs neural rendering of multi-view images while utilizing surface normals estimated by a deep photometric stereo network. We render the MVPS images by considering the object's surface normals for each 3D sample point along the viewing direction rather than explicitly using the density gradient in the volume space via 3D occupancy information. We optimize the proposed neural radiance field representation for the MVPS setup efficiently using a fully connected deep network to recover the 3D geometry of an object. Extensive evaluation on the DiLiGenT-MV benchmark dataset shows that our method performs better than the approaches that perform only PS or only multi-view stereo (MVS) and provides comparable results against the state-of-the-art multi-stage fusion methods.

(2 min) Temporal language grounding in videos aims to localize the temporal span relevant to the given query sentence. Previous methods treat it either as a boundary regression task or a span extraction task. This paper will formulate temporal language grounding into video reading comprehension and propose a Relation-aware Network (RaNet) to address it. This framework aims to select a video moment choice from the predefined answer set with the aid of coarse-and-fine choice-query interaction and choice-choice relation construction. A choice-query interactor is proposed to match the visual and textual information simultaneously in sentence-moment and token-moment levels, leading to a coarse-and-fine cross-modal interaction. Moreover, a novel multi-choice relation constructor is introduced by leveraging graph convolution to capture the dependencies among video moment choices for the best choice selection. Extensive experiments on ActivityNet-Captions, TACoS, and Charades-STA demonstrate the effectiveness of our solution. Codes will be released soon.

(0 min) Understanding the training dynamics of deep learning models is perhaps a necessary step toward demystifying the effectiveness of these models. In particular, how do data from different classes gradually become separable in their feature spaces when training neural networks using stochastic gradient descent? In this study, we model the evolution of features during deep learning training using a set of stochastic differential equations (SDEs) that each corresponds to a training sample. As a crucial ingredient in our modeling strategy, each SDE contains a drift term that reflects the impact of backpropagation at an input on the features of all samples. Our main finding uncovers a sharp phase transition phenomenon regarding the {intra-class impact: if the SDEs are locally elastic in the sense that the impact is more significant on samples from the same class as the input, the features of the training data become linearly separable, meaning vanishing training loss; otherwise, the features are not separable, regardless of how long the training time is. Moreover, in the presence of local elasticity, an analysis of our SDEs shows that the emergence of a simple geometric structure called the neural collapse of the features. Taken together, our results shed light on the decisive role of local elasticity in the training dynamics of neural networks. We corroborate our theoretical analysis with experiments on a synthesized dataset of geometric shapes and CIFAR-10.

(0 min) Inspired by diversity of biological neurons, quadratic artificial neurons can play an important role in deep learning models. The type of quadratic neurons of our interest replaces the inner-product operation in the conventional neuron with a quadratic function. Despite promising results so far achieved by networks of quadratic neurons, there are important issues not well addressed. Theoretically, the superior expressivity of a quadratic network over either a conventional network or a conventional network via quadratic activation is not fully elucidated, which makes the use of quadratic networks not well grounded. Practically, although a quadratic network can be trained via generic backpropagation, it can be subject to a higher risk of collapse than the conventional counterpart. To address these issues, we first apply the spline theory and a measure from algebraic geometry to give two theorems that demonstrate better model expressivity of a quadratic network than the conventional counterpart with or without quadratic activation. Then, we propose an effective and efficient training strategy referred to as ReLinear to stabilize the training process of a quadratic network, thereby unleashing the full potential in its associated machine learning tasks. Comprehensive experiments on popular datasets are performed to support our findings and evaluate the performance of quadratic deep learning.

(2 min) Background & Aims: Hepatic steatosis is a major cause of chronic liver disease. 2D ultrasound is the most widely used non-invasive tool for screening and monitoring, but associated diagnoses are highly subjective. We developed a scalable deep learning (DL) algorithm for quantitative scoring of liver steatosis from 2D ultrasound images. Approach & Results: Using retrospectively collected multi-view ultrasound data from 3,310 patients, 19,513 studies, and 228,075 images, we trained a DL algorithm to diagnose steatosis stages (healthy, mild, moderate, or severe) from ultrasound diagnoses. Performance was validated on two multi-scanner unblinded and blinded (initially to DL developer) histology-proven cohorts (147 and 112 patients) with histopathology fatty cell percentage diagnoses, and a subset with FibroScan diagnoses. We also quantified reliability across scanners and viewpoints. Results were evaluated using Bland-Altman and receiver operating characteristic (ROC) analysis. The DL algorithm demonstrates repeatable measurements with a moderate number of images (3 for each viewpoint) and high agreement across 3 premium ultrasound scanners. High diagnostic performance was observed across all viewpoints: area under the curves of the ROC to classify >=mild, >=moderate, =severe steatosis grades were 0.85, 0.90, and 0.93, respectively. The DL algorithm outperformed or performed at least comparably to FibroScan with statistically significant improvements for all levels on the unblinded histology-proven cohort, and for =severe steatosis on the blinded histology-proven cohort. Conclusions: The DL algorithm provides a reliable quantitative steatosis assessment across view and scanners on two multi-scanner cohorts. Diagnostic performance was high with comparable or better performance than FibroScan.

(2 min) We present an automated machine learning approach for uncalibrated photometric stereo (PS). Our work aims at discovering lightweight and computationally efficient PS neural networks with excellent surface normal accuracy. Unlike previous uncalibrated deep PS networks, which are handcrafted and carefully tuned, we leverage differentiable neural architecture search (NAS) strategy to find uncalibrated PS architecture automatically. We begin by defining a discrete search space for a light calibration network and a normal estimation network, respectively. We then perform a continuous relaxation of this search space and present a gradient-based optimization strategy to find an efficient light calibration and normal estimation network. Directly applying the NAS methodology to uncalibrated PS is not straightforward as certain task-specific constraints must be satisfied, which we impose explicitly. Moreover, we search for and train the two networks separately to account for the Generalized Bas-Relief (GBR) ambiguity. Extensive experiments on the DiLiGenT dataset show that the automatically searched neural architectures performance compares favorably with the state-of-the-art uncalibrated PS methods while having a lower memory footprint.

(2 min) Robust principal component analysis (RPCA) is a critical tool in modern machine learning, which detects outliers in the task of low-rank matrix reconstruction. In this paper, we propose a scalable and learnable non-convex approach for high-dimensional RPCA problems, which we call Learned Robust PCA (LRPCA). LRPCA is highly efficient, and its free parameters can be effectively learned to optimize via deep unfolding. Moreover, we extend deep unfolding from finite iterations to infinite iterations via a novel feedforward-recurrent-mixed neural network model. We establish the recovery guarantee of LRPCA under mild assumptions for RPCA. Numerical experiments show that LRPCA outperforms the state-of-the-art RPCA algorithms, such as ScaledGD and AltProj, on both synthetic datasets and real-world applications.

(2 min) The advancement of technology has progressed faster than any other field in the world and with the development of these new technologies, it is important to make sure that these tools can be used by everyone, including people with disabilities. Accessibility options in computing devices help ensure that everyone has the same access to advanced technologies. Unfortunately, for those who require more unique and sometimes challenging accommodations, such as people with Amyotrophic lateral sclerosis ( ALS), the most commonly used accessibility features are simply not enough. While assistive technology for those with ALS does exist, it requires multiple peripheral devices that can become quite expensive collectively. The purpose of this paper is to suggest a more affordable and readily available option for ALS assistive technology that can be implemented on a smartphone or tablet.

(2 min) Deep neural networks are known to be vulnerable to adversarially perturbed inputs. A commonly used defense is adversarial training, whose performance is influenced by model capacity. While previous works have studied the impact of varying model width and depth on robustness, the impact of increasing capacity by using learnable parametric activation functions (PAFs) has not been studied. We study how using learnable PAFs can improve robustness in conjunction with adversarial training. We first ask the question: how should we incorporate parameters into activation functions to improve robustness? To address this, we analyze the direct impact of activation shape on robustness through PAFs and observe that activation shapes with positive outputs on negative inputs and with high finite curvature can increase robustness. We combine these properties to create a new PAF, which we call Parametric Shifted Sigmoidal Linear Unit (PSSiLU). We then combine PAFs (including PReLU, PSoftplus and PSSiLU) with adversarial training and analyze robust performance. We find that PAFs optimize towards activation shape properties found to directly affect robustness. Additionally, we find that while introducing only 1-2 learnable parameters into the network, smooth PAFs can significantly increase robustness over ReLU. For instance, when trained on CIFAR-10 with additional synthetic data, PSSiLU improves robust accuracy by 4.54% over ReLU on ResNet-18 and 2.69% over ReLU on WRN-28-10 in the $\ell_{\infty}$ threat model while adding only 2 additional parameters into the network architecture. The PSSiLU WRN-28-10 model achieves 61.96% AutoAttack accuracy, improving over the state-of-the-art robust accuracy on RobustBench (Croce et al., 2020).

(2 min) Most existing neural architecture search (NAS) benchmarks and algorithms prioritize performance on well-studied tasks, e.g., image classification on CIFAR and ImageNet. This makes the applicability of NAS approaches in more diverse areas inadequately understood. In this paper, we present NAS-Bench-360, a benchmark suite for evaluating state-of-the-art NAS methods for convolutional neural networks (CNNs). To construct it, we curate a collection of ten tasks spanning a diverse array of application domains, dataset sizes, problem dimensionalities, and learning objectives. By carefully selecting tasks that can both interoperate with modern CNN-based search methods but that are also far-afield from their original development domain, we can use NAS-Bench-360 to investigate the following central question: do existing state-of-the-art NAS methods perform well on diverse tasks? Our experiments show that a modern NAS procedure designed for image classification can indeed find good architectures for tasks with other dimensionalities and learning objectives; however, the same method struggles against more task-specific methods and performs catastrophically poorly on classification in non-vision domains. The case for NAS robustness becomes even more dire in a resource-constrained setting, where a recent NAS method provides little-to-no benefit over much simpler baselines. These results demonstrate the need for a benchmark such as NAS-Bench-360 to help develop NAS approaches that work well on a variety of tasks, a crucial component of a truly robust and automated pipeline. We conclude with a demonstration of the kind of future research our suite of tasks will enable. All data and code is made publicly available.

(2 min) We present a method that takes as input a single dual-pixel image, and simultaneously estimates the image's defocus map -- the amount of defocus blur at each pixel -- and recovers an all-in-focus image. Our method is inspired from recent works that leverage the dual-pixel sensors available in many consumer cameras to assist with autofocus, and use them for recovery of defocus maps or all-in-focus images. These prior works have solved the two recovery problems independently of each other, and often require large labeled datasets for supervised training. By contrast, we show that it is beneficial to treat these two closely-connected problems simultaneously. To this end, we set up an optimization problem that, by carefully modeling the optics of dual-pixel images, jointly solves both problems. We use data captured with a consumer smartphone camera to demonstrate that, after a one-time calibration step, our approach improves upon prior works for both defocus map estimation and blur removal, despite being entirely unsupervised.

(2 min) Recognising previously visited locations is an important, but unsolved, task in autonomous navigation. Current visual place recognition (VPR) benchmarks typically challenge models to recover the position of a query image (or images) from sequential datasets that include both spatial and temporal components. Recently, Echo State Network (ESN) varieties have proven particularly powerful at solving machine learning tasks that require spatio-temporal modelling. These networks are simple, yet powerful neural architectures that -- exhibiting memory over multiple time-scales and non-linear high-dimensional representations -- can discover temporal relations in the data while still maintaining linearity in the learning. In this paper, we present a series of ESNs and analyse their applicability to the VPR problem. We report that the addition of ESNs to pre-processed convolutional neural networks led to a dramatic boost in performance in comparison to non-recurrent networks in four standard benchmarks (GardensPoint, SPEDTest, ESSEX3IN1, Nordland) demonstrating that ESNs are able to capture the temporal structure inherent in VPR problems. Moreover, we show that ESNs can outperform class-leading VPR models which also exploit the sequential dynamics of the data. Finally, our results demonstrate that ESNs also improve generalisation abilities, robustness, and accuracy further supporting their suitability to VPR applications.

(2 min) Relying on monocular image data for precise 3D object detection remains an open problem, whose solution has broad implications for cost-sensitive applications such as traffic monitoring. We present UrbanNet, a modular architecture for long range monocular 3D object detection with static cameras. Our proposed system combines commonly available urban maps along with a mature 2D object detector and an efficient 3D object descriptor to accomplish accurate detection at long range even when objects are rotated along any of their three axes. We evaluate UrbanNet on a novel challenging synthetic dataset and highlight the advantages of its design for traffic detection in roads with changing slope, where the flat ground approximation does not hold. Data and code are available at https://github.com/TRAILab/UrbanNet

(2 min) Training machines to understand natural language and interact with humans is an elusive and essential task of artificial intelligence. A diversity of dialogue systems has been designed with the rapid development of deep learning techniques, especially the recent pre-trained language models (PrLMs). Among these studies, the fundamental yet challenging type of task is dialogue comprehension whose role is to teach the machines to read and comprehend the dialogue context before responding. In this paper, we review the previous methods from the technical perspective of dialogue modeling for the dialogue comprehension task. We summarize the characteristics and challenges of dialogue comprehension in contrast to plain-text reading comprehension. Then, we discuss three typical patterns of dialogue modeling. In addition, we categorize dialogue-related pre-training techniques which are employed to enhance PrLMs in dialogue scenarios. Finally, we highlight the technical advances in recent years and point out the lessons from the empirical analysis and the prospects towards a new frontier of researches.

(2 min) Performance of recommender systems (RS) relies heavily on the amount of training data available. This poses a chicken-and-egg problem for early-stage products, whose amount of data, in turn, relies on the performance of their RS. On the other hand, zero-shot learning promises some degree of generalization from an old dataset to an entirely new dataset. In this paper, we explore the possibility of zero-shot learning in RS. We develop an algorithm, dubbed ZEro-Shot Recommenders (ZESRec), that is trained on an old dataset and generalize to a new one where there are neither overlapping users nor overlapping items, a setting that contrasts typical cross-domain RS that has either overlapping users or items. Different from categorical item indices, i.e., item ID, in previous methods, ZESRec uses items' natural-language descriptions (or description embeddings) as their continuous indices, and therefore naturally generalize to any unseen items. In terms of users, ZESRec builds upon recent advances on sequential RS to represent users using their interactions with items, thereby generalizing to unseen users as well. We study three pairs of real-world RS datasets and demonstrate that ZESRec can successfully enable recommendations in such a zero-shot setting, opening up new opportunities for resolving the chicken-and-egg problem for data-scarce startups or early-stage products.

(2 min) In contrast to traditional online videos, live multi-streaming supports real-time social interactions between multiple streamers and viewers, such as donations. However, donation and multi-streaming channel recommendations are challenging due to complicated streamer and viewer relations, asymmetric communications, and the tradeoff between personal interests and group interactions. In this paper, we introduce Multi-Stream Party (MSP) and formulate a new multi-streaming recommendation problem, called Donation and MSP Recommendation (DAMRec). We propose Multi-stream Party Recommender System (MARS) to extract latent features via socio-temporal coupled donation-response tensor factorization for donation and MSP recommendations. Experimental results on Twitch and Douyu manifest that MARS significantly outperforms existing recommenders by at least 38.8% in terms of hit ratio and mean average precision.

(2 min) Racial disparity in academia is a widely acknowledged problem. The quantitative understanding of racial based systemic inequalities is an important step towards a more equitable research system. However, because of the lack of robust information on authors' race, few large scale analyses have been performed on this topic. Algorithmic approaches offer one solution, using known information about authors, such as their names, to infer their perceived race. As with any other algorithm, the process of racial inference can generate biases if it is not carefully considered. The goal of this article is to assess the extent to which algorithmic bias is introduced using different approaches for name based racial inference. We use information from the U.S. Census and mortgage applications to infer the race of U.S. affiliated authors in the Web of Science. We estimate the effects of using given and family names, thresholds or continuous distributions, and imputation. Our results demonstrate that the validity of name based inference varies by race/ethnicity and that threshold approaches underestimate Black authors and overestimate White authors. We conclude with recommendations to avoid potential biases. This article lays the foundation for more systematic and less biased investigations into racial disparities in science.

(2 min) Recent advancements of sequential deep learning models such as Transformer and BERT have significantly facilitated the sequential recommendation. However, according to our study, the distribution of item embeddings generated by these models tends to degenerate into an anisotropic shape, which may result in high semantic similarities among embeddings. In this paper, both empirical and theoretical investigations of this representation degeneration problem are first provided, based on which a novel recommender model DuoRec is proposed to improve the item embeddings distribution. Specifically, in light of the uniformity property of contrastive learning, a contrastive regularization is designed for DuoRec to reshape the distribution of sequence representations. Given the convention that the recommendation task is performed by measuring the similarity between sequence representations and item embeddings in the same space via dot product, the regularization can be implicitly applied to the item embedding distribution. Existing contrastive learning methods mainly rely on data level augmentation for user-item interaction sequences through item cropping, masking, or reordering and can hardly provide semantically consistent augmentation samples. In DuoRec, a model-level augmentation is proposed based on Dropout to enable better semantic preserving. Furthermore, a novel sampling strategy is developed, where sequences having the same target item are chosen hard positive samples. Extensive experiments conducted on five datasets demonstrate the superior performance of the proposed DuoRec model compared with baseline methods. Visualization results of the learned representations validate that DuoRec can largely alleviate the representation degeneration problem.

(2 min) We present principled approaches to train and deploy dyadic neural embedding models at the billion scale, focusing our investigation on the application of semantic product search. When training a dyadic model, one seeks to embed two different types of entities (e.g., queries and documents or users and movies) in a common vector space such that pairs with high relevance are positioned nearby. During inference, given an embedding of one type (e.g., a query or a user), one seeks to retrieve the entities of the other type (e.g., documents or movies, respectively) that are highly relevant. In this work, we show that exploiting the natural structure of real-world datasets helps address both challenges efficiently. Specifically, we model dyadic data as a bipartite graph with edges between pairs with positive associations. We then propose to partition this network into semantically coherent clusters and thus reduce our search space by focusing on a small subset of these partitions for a given input. During training, this technique enables us to efficiently mine hard negative examples while, at inference, we can quickly find the nearest neighbors for a given embedding. We provide offline experimental results that demonstrate the efficacy of our techniques for both training and inference on a billion-scale Amazon.com product search dataset.

(2 min) Topline hotels are now shifting into the digital way in how they understand their customers to maintain and ensuring satisfaction. Rather than the conventional way which uses written reviews or interviews, the hotel is now heavily investing in Artificial Intelligence particularly Machine Learning solutions. Analysis of online customer reviews changes the way companies make decisions in a more effective way than using conventional analysis. The purpose of this research is to measure hotel service quality. The proposed approach emphasizes service quality dimensions reviews of the top-5 luxury hotel in Indonesia that appear on the online travel site TripAdvisor based on section Best of 2018. In this research, we use a model based on a simple Bayesian classifier to classify each customer review into one of the service quality dimensions. Our model was able to separate each classification properly by accuracy, kappa, recall, precision, and F-measure measurements. To uncover latent topics in the customer's opinion we use Topic Modeling. We found that the common issue that occurs is about responsiveness as it got the lowest percentage compared to others. Our research provides a faster outlook of hotel rank based on service quality to end customers based on a summary of the previous online review.

(2 min) A recommender system learns to predict the user-specific preference or intention over many items simultaneously for all users, making personalized recommendations based on a relatively small number of observations. One central issue is how to leverage three-way interactions, referred to as user-item-stage dependencies on a monotonic chain of events, to enhance the prediction accuracy. A monotonic chain of events occurs, for instance, in an article sharing dataset, where a ``follow'' action implies a ``like'' action, which in turn implies a ``view'' action. In this article, we develop a multistage recommender system utilizing a two-level monotonic property characterizing a monotonic chain of events for personalized prediction. Particularly, we derive a large-margin classifier based on a nonnegative additive latent factor model in the presence of a high percentage of missing observations, particularly between stages, reducing the number of model parameters for personalized prediction while guaranteeing prediction consistency. On this ground, we derive a regularized cost function to learn user-specific behaviors at different stages, linking decision functions to numerical and categorical covariates to model user-item-stage interactions. Computationally, we derive an algorithm based on blockwise coordinate descent. Theoretically, we show that the two-level monotonic property enhances the accuracy of learning as compared to a standard method treating each stage individually and an ordinal method utilizing only one-level monotonicity. Finally, the proposed method compares favorably with existing methods in simulations and an article sharing dataset.

(2 min) Twitter is a social network that offers a rich and interesting source of information challenging to retrieve and analyze. Twitter data can be accessed using a REST API. The available operations allow retrieving tweets on the basis of a set of keywords but with limitations such as the number of calls per minute and the size of results. Besides, there is no control on retrieved results and finding tweets which are relevant to a specific topic is a big issue. Given these limitations, it is important that the query keywords cover unambiguously the topic of interest in order to both reach the relevant answers and decrease the number of API calls. In this paper, we introduce a new crawling algorithm called "SmartTwitter Crawling" (STiC) that retrieves a set of tweets related to a target topic. In this algorithm, we take an initial keyword query and enrich it using a set of additional keywords that come from different data sources. STiC algorithm relies on a DFS search in Twittergraph where each reached tweet is considered if it is relevant with the query keywords using a scoring, updated throughout the whole crawling process. This scoring takes into account the tweet text, hashtags and the users who have posted the tweet, replied to the tweet, been mentioned in the tweet or retweeted the tweet. Given this score, STiC is able to select relevant tweets in each iteration and continue by adding the related valuable tweets. Several experiments have been achieved for different kinds of queries, the results showedthat the precision increases compared to a simple BFS search.

(2 min) Ranking system is the core part of modern retrieval and recommender systems, where the goal is to rank candidate items given user contexts. Optimizing ranking systems online means that the deployed system can serve user requests, e.g., queries in the web search, and optimize the ranking policy by learning from user interactions, e.g., clicks. Bandit is a general online learning framework and can be used in our optimization task. However, due to the unique features of ranking, there are several challenges in designing bandit algorithms for ranking system optimization. In this dissertation, we study and propose solutions for four challenges in optimizing ranking systems online: effectiveness, safety, nonstationarity, and diversification. First, the effectiveness is related to how fast the algorithm learns from interactions. We study the effective online ranker evaluation task and propose the MergeDTS algorithm to solve the problem effectively. Second, the deployed algorithm should be safe, which means the algorithm only displays reasonable content to user requests. To solve the safe online learning to rank problem, we propose the BubbleRank algorithm. Third, as users change their preferences constantly, the algorithm should handle the nonstationarity. We formulate this nonstationary online learning to rank problem as cascade non-stationary bandits and propose CascadeDUCB and CascadeSWUCB algorithms to solve the problem. Finally, the contents in ranked lists should be diverse. We consider the results diversification task and propose the CascadeHybird algorithm that considers both the item relevance and results diversification when learning from user interactions.

(2 min) Neural approaches, specifically transformer models, for ranking documents have delivered impressive gains in ranking performance. However, query processing using such over-parameterized models is both resource and time intensive. Consequently, to keep query processing costs manageable, trade-offs are made to reduce the number of documents to be re-ranked or consider leaner models with fewer parameters. In this paper, we propose the fast-forward index -- a simple vector forward index that facilitates ranking documents using interpolation-based ranking models. Fast-forward indexes pre-compute the dense transformer-based vector representations of documents and passages for fast CPU-based semantic similarity computation during query processing. We propose theoretically grounded index pruning and early stopping techniques to improve the query-processing throughput using fast-forward indexes. We conduct extensive large-scale experiments over the TREC-DL datasets and show up to 75% improvement in query-processing performance over hybrid indexes using only CPUs. Along with the efficiency benefits, we show that fast-forward indexes can deliver superior ranking performance due to the complementary benefits of interpolation between lexical and semantic similarities.

(2 min) Dense Retrieval (DR) has achieved state-of-the-art first-stage ranking effectiveness. However, the efficiency of most existing DR models is limited by the large memory cost of storing dense vectors and the time-consuming nearest neighbor search (NNS) in vector space. Therefore, we present RepCONC, a novel retrieval model that learns discrete Representations via CONstrained Clustering. RepCONC jointly trains dual-encoders and the Product Quantization (PQ) method to learn discrete document representations and enables fast approximate NNS with compact indexes. It models quantization as a constrained clustering process, which requires the document embeddings to be uniformly clustered around the quantization centroids and supports end-to-end optimization of the quantization method and dual-encoders. We theoretically demonstrate the importance of the uniform clustering constraint in RepCONC and derive an efficient approximate solution for constrained clustering by reducing it to an instance of the optimal transport problem. Besides constrained clustering, RepCONC further adopts a vector-based inverted file system (IVF) to support highly efficient vector search on CPUs. Extensive experiments on two popular ad-hoc retrieval benchmarks show that RepCONC achieves better ranking effectiveness than competitive vector quantization baselines under different compression ratio settings. It also substantially outperforms a wide range of existing retrieval models in terms of retrieval effectiveness, memory efficiency, and time efficiency.

(2 min) Controllable music generation with deep generative models has become increasingly reliant on disentanglement learning techniques. However, current disentanglement metrics, such as mutual information gap (MIG), are often inadequate and misleading when used for evaluating latent representations in the presence of interdependent semantic attributes often encountered in real-world music datasets. In this work, we propose a dependency-aware information metric as a drop-in replacement for MIG that accounts for the inherent relationship between semantic attributes.

(2 min) News recommender systems are essential for helping users to efficiently and effectively find out those interesting news from a large amount of news. Most of existing news recommender systems usually learn topic-level representations of users and news for recommendation, and neglect to learn more informative aspect-level features of users and news for more accurate recommendation. As a result, they achieve limited recommendation performance. Aiming at addressing this deficiency, we propose a novel Aspect-driven News Recommender System (ANRS) built on aspect-level user preference and news representation learning. Here, \textit{news aspect} is fine-grained semantic information expressed by a set of related words, which indicates specific aspects described by the news. In ANRS, \textit{news aspect-level encoder} and \textit{user aspect-level encoder} are devised to learn the fine-grained aspect-level representations of user's preferences and news characteristics respectively, which are fed into \textit{click predictor} to judge the probability of the user clicking the candidate news. Extensive experiments are done on the commonly used real-world dataset MIND, which demonstrate the superiority of our method compared with representative and state-of-the-art methods.

(2 min) We describe our workflow to create an engaging remote learning experience for a university course, while minimizing the post-production time of the educators. We make use of ubiquitous and commonly free services and platforms, so that our workflow is inclusive for all educators and provides polished experiences for students. Our learning materials provide for each lecture: 1) a recorded video, uploaded on YouTube, with exact slide timestamp indices, which enables an enhanced navigation UI; and 2) a high-quality flow-text automated transcript of the narration with proper punctuation and capitalization, improved with a student participation workflow on GitHub. All these results could be created by hand in a time consuming and costly way. However, this would generally exceed the time available for creating course materials. Our main contribution is to automate the transformation and post-production between raw narrated slides and our published materials with a custom toolchain. Furthermore, we describe our complete workflow: from content creation to transformation and distribution. Our students gave us overwhelmingly positive feedback and especially liked our use of ubiquitous platforms. The most used feature was YouTube's chapter UI enabled through our automatically generated timestamps. The majority of students, who started using the transcripts, continued to do so. Every single transcript was corrected by students, with an average word-change of 6%. We conclude with the positive feedback that our enhanced content formats are much appreciated and utilized. Important for educators is how our low overhead production workflow was sustainable throughout a busy semester.

(2 min) Tremendous progress has been made in sequential processing with the recent advances in recurrent neural networks. However, recurrent architectures face the challenge of exploding/vanishing gradients during training, and require significant computational resources to execute back-propagation through time. Moreover, large models are typically needed for executing complex sequential tasks. To address these challenges, we propose a novel neuron model that has cosine activation with a time varying component for sequential processing. The proposed neuron provides an efficient building block for projecting sequential inputs into spectral domain, which helps to retain long-term dependencies with minimal extra model parameters and computation. A new type of recurrent network architecture, named Oscillatory Fourier Neural Network, based on the proposed neuron is presented and applied to various types of sequential tasks. We demonstrate that recurrent neural network with the proposed neuron model is mathematically equivalent to a simplified form of discrete Fourier transform applied onto periodical activation. In particular, the computationally intensive back-propagation through time in training is eliminated, leading to faster training while achieving the state of the art inference accuracy in a diverse group of sequential tasks. For instance, applying the proposed model to sentiment analysis on IMDB review dataset reaches 89.4% test accuracy within 5 epochs, accompanied by over 35x reduction in the model size compared to LSTM. The proposed novel RNN architecture is well poised for intelligent sequential processing in resource constrained hardware.

(2 min) Human activity recognition (HAR) is an important research field in ubiquitous computing where the acquisition of large-scale labeled sensor data is tedious, labor-intensive and time consuming. State-of-the-art unsupervised remedies investigated to alleviate the burdens of data annotations in HAR mainly explore training autoencoder frameworks. In this paper: we explore generative adversarial network (GAN) paradigms to learn unsupervised feature representations from wearable sensor data; and design a new GAN framework-Geometrically-Guided GAN or Guided-GAN-for the task. To demonstrate the effectiveness of our formulation, we evaluate the features learned by Guided-GAN in an unsupervised manner on three downstream classification benchmarks. Our results demonstrate Guided-GAN to outperform existing unsupervised approaches whilst closely approaching the performance with fully supervised learned representations. The proposed approach paves the way to bridge the gap between unsupervised and supervised human activity recognition whilst helping to reduce the cost of human data annotation tasks.

(2 min) The prototypical network is a prototype classifier based on meta-learning and is widely used for few-shot learning because it classifies unseen examples by constructing class-specific prototypes without adjusting hyper-parameters during meta-testing. Interestingly, recent research has attracted a lot of attention, showing that a linear classifier with fine-tuning, which does not use a meta-learning algorithm, performs comparably with the prototypical network. However, fine-tuning requires additional hyper-parameters when adapting a model to a new environment. In addition, although the purpose of few-shot learning is to enable the model to quickly adapt to a new environment, fine-tuning needs to be applied every time a new class appears, making fast adaptation difficult. In this paper, we analyze how a prototype classifier works equally well without fine-tuning and meta-learning. We experimentally found that directly using the feature vector extracted using standard pre-trained models to construct a prototype classifier in meta-testing does not perform as well as the prototypical network and linear classifiers with fine-tuning and feature vectors of pre-trained models. Thus, we derive a novel generalization bound for the prototypical network and show that focusing on the variance of the norm of a feature vector can improve performance. We experimentally investigated several normalization methods for minimizing the variance of the norm and found that the same performance can be obtained by using the L2 normalization and embedding space transformation without fine-tuning or meta-learning.

(2 min) Physics-informed neural networks (PINNs) have become a popular choice for solving high-dimensional partial differential equations (PDEs) due to their excellent approximation power and generalization ability. Recently, Extended PINNs (XPINNs) based on domain decomposition methods have attracted considerable attention due to their effectiveness in modeling multiscale and multiphysics problems and their parallelization. However, theoretical understanding on their convergence and generalization properties remains unexplored. In this study, we take an initial step towards understanding how and when XPINNs outperform PINNs. Specifically, for general multi-layer PINNs and XPINNs, we first provide a prior generalization bound via the complexity of the target functions in the PDE problem, and a posterior generalization bound via the posterior matrix norms of the networks after optimization. Moreover, based on our bounds, we analyze the conditions under which XPINNs improve generalization. Concretely, our theory shows that the key building block of XPINN, namely the domain decomposition, introduces a tradeoff for generalization. On the one hand, XPINNs decompose the complex PDE solution into several simple parts, which decreases the complexity needed to learn each part and boosts generalization. On the other hand, decomposition leads to less training data being available in each subdomain, and hence such model is typically prone to overfitting and may become less generalizable. Empirically, we choose five PDEs to show when XPINNs perform better than, similar to, or worse than PINNs, hence demonstrating and justifying our new theory.

(2 min) Deep neural networks (DNNs) with the flexibility to learn good top-layer representations have eclipsed shallow kernel methods without that flexibility. Here, we take inspiration from DNNs to develop the first non-Bayesian deep kernel method, the deep kernel machine. In addition, we develop a solver for the intermediate layer kernels in deep kernel machines that converges in around 10 steps, exploiting matrix solvers initially developed in the control theory literature. These are many times faster the usual gradient descent approach and generalise to arbitrary architectures. While deep kernel machines currently scale poorly in the number of datapoints, we believe that this can be rectified in future work, allowing deep kernel machines to form the basis of a new class of much more efficient deep nonlinear function approximators.

(2 min) Real world learning scenarios involve a nonstationary distribution of classes with sequential dependencies among the samples, in contrast to the standard machine learning formulation of drawing samples independently from a fixed, typically uniform distribution. Furthermore, real world interactions demand learning on-the-fly from few or no class labels. In this work, we propose an unsupervised model that simultaneously performs online visual representation learning and few-shot learning of new categories without relying on any class labels. Our model is a prototype-based memory network with a control component that determines when to form a new class prototype. We formulate it as an online Gaussian mixture model, where components are created online with only a single new example, and assignments do not have to be balanced, which permits an approximation to natural imbalanced distributions from uncurated raw data. Learning includes a contrastive loss that encourages different views of the same image to be assigned to the same prototype. The result is a mechanism that forms categorical representations of objects in nonstationary environments. Experiments show that our method can learn from an online stream of visual input data and is significantly better at category recognition compared to state-of-the-art self-supervised learning methods.

(2 min) Multi-modal reasoning systems rely on a pre-trained object detector to extract regions of interest from the image. However, this crucial module is typically used as a black box, trained independently of the downstream task and on a fixed vocabulary of objects and attributes. This makes it challenging for such systems to capture the long tail of visual concepts expressed in free form text. In this paper we propose MDETR, an end-to-end modulated detector that detects objects in an image conditioned on a raw text query, like a caption or a question. We use a transformer-based architecture to reason jointly over text and image by fusing the two modalities at an early stage of the model. We pre-train the network on 1.3M text-image pairs, mined from pre-existing multi-modal datasets having explicit alignment between phrases in text and objects in the image. We then fine-tune on several downstream tasks such as phrase grounding, referring expression comprehension and segmentation, achieving state-of-the-art results on popular benchmarks. We also investigate the utility of our model as an object detector on a given label set when fine-tuned in a few-shot setting. We show that our pre-training approach provides a way to handle the long tail of object categories which have very few labelled instances. Our approach can be easily extended for visual question answering, achieving competitive performance on GQA and CLEVR. The code and models are available at https://github.com/ashkamath/mdetr.

(2 min) Quantified Boolean logic results from adding operators to Boolean logic for existentially and universally quantifying variables. This extends the reach of Boolean logic by enabling a variety of applications that have been explored over the decades. The existential quantification of literals (variable states) and its applications have also been studied in the literature. In this paper, we complement this by studying universal literal quantification and its applications, particularly to explainable AI. We also provide a novel semantics for quantification, discuss the interplay between variable/literal and existential/universal quantification. We further identify some classes of Boolean formulas and circuits on which quantification can be done efficiently. Literal quantification is more fine-grained than variable quantification as the latter can be defined in terms of the former. This leads to a refinement of quantified Boolean logic with literal quantification as its primitive.

(2 min) The energy consumption of deep learning models is increasing at a breathtaking rate, which raises concerns due to potential negative effects on carbon neutrality in the context of global warming and climate change. With the progress of efficient deep learning techniques, e.g., model compression, researchers can obtain efficient models with fewer parameters and smaller latency. However, most of the existing efficient deep learning methods do not explicitly consider energy consumption as a key performance indicator. Furthermore, existing methods mostly focus on the inference costs of the resulting efficient models, but neglect the notable energy consumption throughout the entire life cycle of the algorithm. In this paper, we present the first large-scale energy consumption benchmark for efficient computer vision models, where a new metric is proposed to explicitly evaluate the full-cycle energy consumption under different model usage intensity. The benchmark can provide insights for low carbon emission when selecting efficient deep learning algorithms in different model usage scenarios.

(2 min) The privacy leakage of the model about the training data can be bounded in the differential privacy mechanism. However, for meaningful privacy parameters, a differentially private model degrades the utility drastically when the model comprises a large number of trainable parameters. In this paper, we propose an algorithm \emph{Gradient Embedding Perturbation (GEP)} towards training differentially private deep models with decent accuracy. Specifically, in each gradient descent step, GEP first projects individual private gradient into a non-sensitive anchor subspace, producing a low-dimensional gradient embedding and a small-norm residual gradient. Then, GEP perturbs the low-dimensional embedding and the residual gradient separately according to the privacy budget. Such a decomposition permits a small perturbation variance, which greatly helps to break the dimensional barrier of private learning. With GEP, we achieve decent accuracy with reasonable computational cost and modest privacy guarantee for deep models. Especially, with privacy bound $\epsilon=8$, we achieve $74.9\%$ test accuracy on CIFAR10 and $95.1\%$ test accuracy on SVHN, significantly improving over existing results.

(2 min) It is well noted that coordinate based MLPs benefit -- in terms of preserving high-frequency information -- through the encoding of coordinate positions as an array of Fourier features. Hitherto, the rationale for the effectiveness of these positional encodings has been solely studied through a Fourier lens. In this paper, we strive to broaden this understanding by showing that alternative non-Fourier embedding functions can indeed be used for positional encoding. Moreover, we show that their performance is entirely determined by a trade-off between the stable rank of the embedded matrix and the distance preservation between embedded coordinates. We further establish that the now ubiquitous Fourier feature mapping of position is a special case that fulfills these conditions. Consequently, we present a more general theory to analyze positional encoding in terms of shifted basis functions. To this end, we develop the necessary theoretical formulae and empirically verify that our theoretical claims hold in practice. Codes available at https://github.com/osiriszjq/Rethinking-positional-encoding.

(2 min) Self-supervised representation learning is able to learn semantically meaningful features; however, much of its recent success relies on multiple crops of an image with very few objects. Instead of learning view-invariant representation from simple images, humans learn representations in a complex world with changing scenes by observing object movement, deformation, pose variation, and ego motion. Motivated by this ability, we present a new self-supervised learning representation framework that can be directly deployed on a video stream of complex scenes with many moving objects. Our framework features a simple flow equivariance objective that encourages the network to predict the features of another frame by applying a flow transformation to the features of the current frame. Our representations, learned from high-resolution raw video, can be readily used for downstream tasks on static images. Readout experiments on challenging semantic segmentation, instance segmentation, and object detection benchmarks show that we are able to outperform representations obtained from previous state-of-the-art methods including SimCLR and BYOL.

(2 min) Semantic concepts are frequently defined by combinations of underlying attributes. As mappings from attributes to classes are often simple, attribute-based representations facilitate novel concept learning with zero or few examples. A significant limitation of existing attribute-based learning paradigms, such as zero-shot learning, is that the attributes are assumed to be known and fixed. In this work we study the rapid learning of attributes that were not previously labeled. Compared to standard few-shot learning of semantic classes, in which novel classes may be defined by attributes that were relevant at training time, learning new attributes imposes a stiffer challenge. We found that supervised learning with training attributes does not generalize well to new test attributes, whereas self-supervised pre-training brings significant improvement. We further experimented with random splits of the attribute space and found that predictability of test attributes provides an informative estimate of a model's generalization ability.

(2 min) Complex-valued processing has brought deep learning-based speech enhancement and signal extraction to a new level. Typically, the process is based on a time-frequency (TF) mask which is applied to a noisy spectrogram, while complex masks (CM) are usually preferred over real-valued masks due to their ability to modify the phase. Recent work proposed to use a complex filter instead of a point-wise multiplication with a mask. This allows to incorporate information from previous and future time steps exploiting local correlations within each frequency band. In this work, we propose DeepFilterNet, a two stage speech enhancement framework utilizing deep filtering. First, we enhance the spectral envelope using ERB-scaled gains modeling the human frequency perception. The second stage employs deep filtering to enhance the periodic components of speech. Additionally to taking advantage of perceptual properties of speech, we enforce network sparsity via separable convolutions and extensive grouping in linear and recurrent layers to design a low complexity architecture. We further show that our two stage deep filtering approach outperforms complex masks over a variety of frequency resolutions and latencies and demonstrate convincing performance compared to other state-of-the-art models.

(2 min) This paper presents a predictive model for estimating regularization parameters of diffeomorphic image registration. We introduce a novel framework that automatically determines the parameters controlling the smoothness of diffeomorphic transformations. Our method significantly reduces the effort of parameter tuning, which is time and labor-consuming. To achieve the goal, we develop a predictive model based on deep convolutional neural networks (CNN) that learns the mapping between pairwise images and the regularization parameter of image registration. In contrast to previous methods that estimate such parameters in a high-dimensional image space, our model is built in an efficient bandlimited space with much lower dimensions. We demonstrate the effectiveness of our model on both 2D synthetic data and 3D real brain images. Experimental results show that our model not only predicts appropriate regularization parameters for image registration, but also improving the network training in terms of time and memory efficiency.

(2 min) The study of provable adversarial robustness for deep neural networks (DNNs) has mainly focused on static supervised learning tasks such as image classification. However, DNNs have been used extensively in real-world adaptive tasks such as reinforcement learning (RL), making such systems vulnerable to adversarial attacks as well. Prior works in provable robustness in RL seek to certify the behaviour of the victim policy at every time-step against a non-adaptive adversary using methods developed for the static setting. But in the real world, an RL adversary can infer the defense strategy used by the victim agent by observing the states, actions, etc. from previous time-steps and adapt itself to produce stronger attacks in future steps. We present an efficient procedure, designed specifically to defend against an adaptive RL adversary, that can directly certify the total reward without requiring the policy to be robust at each time-step. Our main theoretical contribution is to prove an adaptive version of the Neyman-Pearson Lemma -- a key lemma for smoothing-based certificates -- where the adversarial perturbation at a particular time can be a stochastic function of current and previous observations and states as well as previous actions. Building on this result, we propose policy smoothing where the agent adds a Gaussian noise to its observation at each time-step before passing it through the policy function. Our robustness certificates guarantee that the final total reward obtained by policy smoothing remains above a certain threshold, even though the actions at intermediate time-steps may change under the attack. Our experiments on various environments like Cartpole, Pong, Freeway and Mountain Car show that our method can yield meaningful robustness guarantees in practice.

(2 min) We study the frontier between learnable and unlearnable hidden Markov models (HMMs). HMMs are flexible tools for clustering dependent data coming from unknown populations. The model parameters are known to be fully identifiable (up to label-switching) without any modeling assumption on the distributions of the populations as soon as the clusters are distinct and the hidden chain is ergodic with a full rank transition matrix. In the limit as any one of these conditions fails, it becomes impossible in general to identify parameters. For a chain with two hidden states we prove nonasymptotic minimax upper and lower bounds, matching up to constants, which exhibit thresholds at which the parameters become learnable. We also provide an upper bound on the relative entropy rate for parameters in a neighbourhood of the unlearnable region which may have interest in itself.

(2 min) Faster inference of deep learning models is highly demanded on edge devices and even servers, for both financial and environmental reasons. To address this issue, we propose SoftNeuro, a novel, high-performance inference framework with efficient performance tuning. The key idea is to separate algorithmic routines from network layers. Our framework maximizes the inference performance by profiling various routines for each layer and selecting the fastest path. To efficiently find the best path, we propose a routine-selection algorithm based on dynamic programming. Experiments show that the proposed framework achieves both fast inference and efficient tuning.

(2 min) The task of learning from only a few examples (called a few-shot setting) is of key importance and relevance to a real-world setting. For question answering (QA), the current state-of-the-art pre-trained models typically need fine-tuning on tens of thousands of examples to obtain good results. Their performance degrades significantly in a few-shot setting (< 100 examples). To address this, we propose a simple fine-tuning framework that leverages pre-trained text-to-text models and is directly aligned with their pre-training framework. Specifically, we construct the input as a concatenation of the question, a mask token representing the answer span and a context. Given this input, the model is fine-tuned using the same objective as that of its pre-training objective. Through experimental studies on various few-shot configurations, we show that this formulation leads to significant gains on multiple QA benchmarks (an absolute gain of 34.2 F1 points on average when there are only 16 training examples). The gains extend further when used with larger models (Eg:- 72.3 F1 on SQuAD using BART-large with only 32 examples) and translate well to a multilingual setting . On the multilingual TydiQA benchmark, our model outperforms the XLM-Roberta-large by an absolute margin of upto 40 F1 points and an average of 33 F1 points in a few-shot setting (<= 64 training examples). We conduct detailed ablation studies to analyze factors contributing to these gains.

(2 min) Stochastic gradient descent (SGD) has been demonstrated to generalize well in many deep learning applications. In practice, one often runs SGD with a geometrically decaying stepsize, i.e., a constant initial stepsize followed by multiple geometric stepsize decay, and uses the last iterate as the output. This kind of SGD is known to be nearly minimax optimal for classical finite-dimensional linear regression problems (Ge et al., 2019), and provably outperforms SGD with polynomially decaying stepsize in terms of the statistical minimax rates. However, a sharp analysis for the last iterate of SGD with decaying step size in the overparameterized setting is still open. In this paper, we provide problem-dependent analysis on the last iterate risk bounds of SGD with decaying stepsize, for (overparameterized) linear regression problems. In particular, for SGD with geometrically decaying stepsize (or tail geometrically decaying stepsize), we prove nearly matching upper and lower bounds on the excess risk. Our results demonstrate the generalization ability of SGD for a wide class of overparameterized problems, and can recover the minimax optimal results up to logarithmic factors in the classical regime. Moreover, we provide an excess risk lower bound for SGD with polynomially decaying stepsize and illustrate the advantage of geometrically decaying stepsize in an instance-wise manner, which complements the minimax rate comparison made in previous work.

(2 min) We introduce the $\gamma$-model, a predictive model of environment dynamics with an infinite probabilistic horizon. Replacing standard single-step models with $\gamma$-models leads to generalizations of the procedures central to model-based control, including the model rollout and model-based value estimation. The $\gamma$-model, trained with a generative reinterpretation of temporal difference learning, is a natural continuous analogue of the successor representation and a hybrid between model-free and model-based mechanisms. Like a value function, it contains information about the long-term future; like a standard predictive model, it is independent of task reward. We instantiate the $\gamma$-model as both a generative adversarial network and normalizing flow, discuss how its training reflects an inescapable tradeoff between training-time and testing-time compounding errors, and empirically investigate its utility for prediction and control.

(2 min) It is well-known that stochastic gradient noise (SGN) acts as implicit regularization for deep learning and is essentially important for both optimization and generalization of deep networks. Some works attempted to artificially simulate SGN by injecting random noise to improve deep learning. However, it turned out that the injected simple random noise cannot work as well as SGN, which is anisotropic and parameter-dependent. For simulating SGN at low computational costs and without changing the learning rate or batch size, we propose the Positive-Negative Momentum (PNM) approach that is a powerful alternative to conventional Momentum in classic optimizers. The introduced PNM method maintains two approximate independent momentum terms. Then, we can control the magnitude of SGN explicitly by adjusting the momentum difference. We theoretically prove the convergence guarantee and the generalization advantage of PNM over Stochastic Gradient Descent (SGD). By incorporating PNM into the two conventional optimizers, SGD with Momentum and Adam, our extensive experiments empirically verified the significant advantage of the PNM-based variants over the corresponding conventional Momentum-based optimizers.

(2 min) Offline reinforcement learning requires reconciling two conflicting aims: learning a policy that improves over the behavior policy that collected the dataset, while at the same time minimizing the deviation from the behavior policy so as to avoid errors due to distributional shift. This trade-off is critical, because most current offline reinforcement learning methods need to query the value of unseen actions during training to improve the policy, and therefore need to either constrain these actions to be in-distribution, or else regularize their values. We propose an offline RL method that never needs to evaluate actions outside of the dataset, but still enables the learned policy to improve substantially over the best behavior in the data through generalization. The main insight in our work is that, instead of evaluating unseen actions from the latest policy, we can approximate the policy improvement step implicitly by treating the state value function as a random variable, with randomness determined by the action (while still integrating over the dynamics to avoid excessive optimism), and then taking a state conditional upper expectile of this random variable to estimate the value of the best actions in that state. This leverages the generalization capacity of the function approximator to estimate the value of the best available action at a given state without ever directly querying a Q-function with this unseen action. Our algorithm alternates between fitting this upper expectile value function and backing it up into a Q-function. Then, we extract the policy via advantage-weighted behavioral cloning. We dub our method implicit Q-learning (IQL). IQL demonstrates the state-of-the-art performance on D4RL, a standard benchmark for offline reinforcement learning. We also demonstrate that IQL achieves strong performance fine-tuning using online interaction after offline initialization.

(2 min) Due the alarming rate of climate change, the implementation of efficient CO$_2$ capture has become crucial. This project aims to create an algorithm that predicts the uptake of CO$_2$ adsorbing Metal-Organic Frameworks (MOFs) by using Machine Learning. These values will in turn gauge the efficiency of these MOFs and provide scientists who are looking to maximize the uptake a way to know whether or not the MOF is worth synthesizing. This algorithm will save resources such as time and equipment as scientists will be able to disregard hypothetical MOFs with low efficiencies. In addition, this paper will also highlight the most important features within the data set. This research will contribute to enable the rapid synthesis of CO$_2$ adsorbing MOFs.

(2 min) Action recognition is a crucial task for video understanding. In this paper, we present AutoVideo, a Python system for automated video action recognition. It currently supports seven action recognition algorithms and various pre-processing modules. Unlike the existing libraries that only provide model zoos, AutoVideo is built with the standard pipeline language. The basic building block is primitive, which wraps a pre-processing module or an algorithm with some hyperparameters. AutoVideo is highly modular and extendable. It can be easily combined with AutoML searchers. The pipeline language is quite general so that we can easily enrich AutoVideo with algorithms for various other video-related tasks in the future. AutoVideo is released under MIT license at https://github.com/datamllab/autovideo

(2 min) Human epidermal growth factor receptor 2 (HER2) is an important prognostic and predictive factor which is overexpressed in 15-20% of breast cancer (BCa). The determination of its status is a key clinical decision making step for selection of treatment regimen and prognostication. HER2 status is evaluated using transcroptomics or immunohistochemistry (IHC) through situ hybridisation (ISH) which require additional costs and tissue burden in addition to analytical variabilities in terms of manual observational biases in scoring. In this study, we propose a novel graph neural network (GNN) based model (termed SlideGraph+) to predict HER2 status directly from whole-slide images of routine Haematoxylin and Eosin (H&E) slides. The network was trained and tested on slides from The Cancer Genome Atlas (TCGA) in addition to two independent test datasets. We demonstrate that the proposed model outperforms the state-of-the-art methods with area under the ROC curve (AUC) values > 0.75 on TCGA and 0.8 on independent test sets. Our experiments show that the proposed approach can be utilised for case triaging as well as pre-ordering diagnostic tests in a diagnostic setting. It can also be used for other weakly supervised prediction problems in computational pathology. The SlideGraph+ code is available at https://github.com/wenqi006/SlideGraph.

(2 min) We investigate two perturbation approaches to overcome conservatism that optimism based algorithms chronically suffer from in practice. The first approach replaces optimism with a simple randomization when using confidence sets. The second one adds random perturbations to its current estimate before maximizing the expected reward. For non-stationary linear bandits, where each action is associated with a $d$-dimensional feature and the unknown parameter is time-varying with total variation $B_T$, we propose two randomized algorithms, Discounted Randomized LinUCB (D-RandLinUCB) and Discounted Linear Thompson Sampling (D-LinTS) via the two perturbation approaches. We highlight the statistical optimality versus computational efficiency trade-off between them in that the former asymptotically achieves the optimal dynamic regret $\tilde{O}(d^{7/8} B_T^{1/4}T^{3/4})$, but the latter is oracle-efficient with an extra logarithmic factor in the number of arms compared to minimax-optimal dynamic regret. In a simulation study, both algorithms show outstanding performance in tackling conservatism issue that Discounted LinUCB struggles with.

(2 min) This paper studies the problem of federated learning (FL) in the absence of a trustworthy server/clients. In this setting, each client needs to ensure the privacy of its own data without relying on the server or other clients. We study local differential privacy (LDP) and provide tight upper and lower bounds that establish the minimax optimal rates (up to logarithms) for LDP convex/strongly convex federated stochastic optimization. Our rates match the optimal statistical rates in certain practical parameter regimes ("privacy for free"). Second, we develop a novel time-varying noisy SGD algorithm, leading to the first non-trivial LDP risk bounds for FL with non-i.i.d. clients. Third, we consider the special case where each client's loss function is empirical and develop an accelerated LDP FL algorithm to improve communication complexity compared to existing works. We also provide matching lower bounds, establishing the optimality of our algorithm for convex/strongly convex settings. Fourth, with a secure shuffler to anonymize client reports (but without a trusted server), our algorithm attains the optimal central DP rates for stochastic convex/strongly convex optimization, thereby achieving optimality in the local and central models simultaneously. Our upper bounds quantify the role of network communication reliability in performance.

(2 min) Hierarchical forecasting is a key problem in many practical multivariate forecasting applications - the goal is to simultaneously predict a large number of correlated time series that are arranged in a pre-specified aggregation hierarchy. The main challenge is to exploit the hierarchical correlations to simultaneously obtain good prediction accuracy for time series at different levels of the hierarchy. In this paper, we propose a new approach for hierarchical forecasting which consists of two components. First, decomposing the time series along a global set of basis time series and modeling hierarchical constraints using the coefficients of the basis decomposition. And second, using a linear autoregressive model with coefficients that vary with time. Unlike past methods, our approach is scalable (inference for a specific time series only needs access to its own history) while also modeling the hierarchical structure via (approximate) coherence constraints among the time series forecasts. We experiment on several public datasets and demonstrate significantly improved overall performance on forecasts at different levels of the hierarchy, compared to existing state-of-the-art hierarchical models.

(2 min) A multi-condition multi-objective optimization method that can find Pareto front over a defined condition space is developed for the first time using deep reinforcement learning. Unlike the conventional methods which perform optimization at a single condition, the present method learns the correlations between conditions and optimal solutions. The exclusive capability of the developed method is examined in the solutions of a novel modified Kursawe benchmark problem and an airfoil shape optimization problem which include nonlinear characteristics which are difficult to resolve using conventional optimization methods. Pareto front with high resolution over a defined condition space is successfully determined in each problem. Compared with multiple operations of a single-condition optimization method for multiple conditions, the present multi-condition optimization method based on deep reinforcement learning shows a greatly accelerated search of Pareto front by reducing the number of required function evaluations. An analysis of aerodynamics performance of airfoils with optimally designed shapes confirms that multi-condition optimization is indispensable to avoid significant degradation of target performance for varying flow conditions.

(2 min) Growing at a fast pace, modern autonomous systems will soon be deployed at scale, opening up the possibility for cooperative multi-agent systems. Sharing information and distributing workloads allow autonomous agents to better perform tasks and increase computation efficiency. However, shared information can be modified to execute adversarial attacks on deep learning models that are widely employed in modern systems. Thus, we aim to study the robustness of such systems and focus on exploring adversarial attacks in a novel multi-agent setting where communication is done through sharing learned intermediate representations of neural networks. We observe that an indistinguishable adversarial message can severely degrade performance, but becomes weaker as the number of benign agents increases. Furthermore, we show that black-box transfer attacks are more difficult in this setting when compared to directly perturbing the inputs, as it is necessary to align the distribution of learned representations with domain adaptation. Our work studies robustness at the neural network level to contribute an additional layer of fault tolerance to modern security protocols for more secure multi-agent systems.

(2 min) Federated learning can be a promising solution for enabling IoT cybersecurity (i.e., anomaly detection in the IoT environment) while preserving data privacy and mitigating the high communication/storage overhead (e.g., high-frequency data from time-series sensors) of centralized over-the-cloud approaches. In this paper, to further push forward this direction with a comprehensive study in both algorithm and system design, we build FedIoT platform that contains FedDetect algorithm for on-device anomaly data detection and a system design for realistic evaluation of federated learning on IoT devices. Furthermore, the proposed FedDetect learning framework improves the performance by utilizing a local adaptive optimizer (e.g., Adam) and a cross-round learning rate scheduler. In a network of realistic IoT devices (Raspberry PI), we evaluate FedIoT platform and FedDetect algorithm in both model and system performance. Our results demonstrate the efficacy of federated learning in detecting a wider range of attack types occurred at multiple devices. The system efficiency analysis indicates that both end-to-end training time and memory cost are affordable and promising for resource-constrained IoT devices. The source code is publicly available at https://github.com/FedML-AI/FedIoT.

(2 min) The diagnosis of cyber-physical systems aims to detect faulty behaviour, its root cause and a mitigation or even prevention policy. Therefore, diagnosis relies on a representation of the system's functional and faulty behaviour combined with observations of the system taken at runtime. The main challenges are the time-intensive building of a model, possible state-explosion while searching for the root cause and interpretability of the results. In this paper we propose a scalable algorithm tackling these challenges. We use a Bayesian network to learn a structured model automatically and optimise the model by a genetic algorithm. Our approach differs from existing work in two aspects: instead of selecting features prior to the analysis we learn a global representation using all available information which is then transformed to a smaller, label-specific one and we focus on interpretability to facilitate repairs. The evaluation shows that our approach is able to learn a model with equal performance to state-of-the-art algorithms while giving better interpretability and having a reduced size.

(2 min) Many efforts have been made for revealing the decision-making process of black-box learning machines such as deep neural networks, resulting in useful local and global explanation methods. For local explanation, stochasticity is known to help: a simple method, called SmoothGrad, has improved the visual quality of gradient-based attribution by adding noise in the input space and taking the average over the noise. In this paper, we extend this idea and propose NoiseGrad that enhances both local and global explanation methods. Specifically, NoiseGrad introduces stochasticity in the weight parameter space, such that the decision boundary is perturbed. NoiseGrad is expected to enhance the local explanation, similarly to SmoothGrad, due to the dual relationship between the input perturbation and the decision boundary perturbation. Furthermore, NoiseGrad can be used to enhance global explanations. We evaluate NoiseGrad and its fusion with SmoothGrad -- FusionGrad -- qualitatively and quantitatively with several evaluation criteria, and show that our novel approach significantly outperforms the baseline methods. Both NoiseGrad and FusionGrad are method-agnostic and as handy as SmoothGrad using simple heuristics for the choice of hyperparameter setting without the need of fine-tuning.

(2 min) Neural Architecture Search (NAS) represents an emerging machine learning (ML) paradigm that automatically searches for models tailored to given tasks, which greatly simplifies the development of ML systems and propels the trend of ML democratization. Yet, little is known about the potential security risks incurred by NAS, which is concerning given the increasing use of NAS-generated models in critical domains. This work represents a solid initial step towards bridging the gap. Through an extensive empirical study of 10 popular NAS methods, we show that compared with their manually designed counterparts, NAS-generated models tend to suffer greater vulnerability to various malicious attacks (e.g., adversarial evasion, model poisoning, and functionality stealing). Further, with both empirical and analytical evidence, we provide possible explanations for such phenomena: given the prohibitive search space and training cost, most NAS methods favor models that converge fast at early training stages; this preference results in architectural properties associated with attack vulnerability (e.g., high loss smoothness and low gradient variance). Our findings not only reveal the relationships between model characteristics and attack vulnerability but also suggest the inherent connections underlying different attacks. Finally, we discuss potential remedies to mitigate such drawbacks, including increasing cell depth and suppressing skip connects, which lead to several promising research directions.

(2 min) Reinforcement Learning (RL) can be considered as a sequence modeling task, i.e., given a sequence of past state-action-reward experiences, a model autoregressively predicts a sequence of future actions. Recently, Transformers have been successfully adopted to model this problem. In this work, we propose State-Action-Reward Transformer (StARformer), which explicitly models local causal relations to help improve action prediction in long sequences. StARformer first extracts local representations (i.e., StAR-representations) from each group of state-action-reward tokens within a very short time span. A sequence of such local representations combined with state representations, is then used to make action predictions over a long time span. Our experiments show that StARformer outperforms the state-of-the-art Transformer-based method on Atari (image) and Gym (state vector) benchmarks, in both offline-RL and imitation learning settings. StARformer is also more compliant with longer sequences of inputs compared to the baseline. Our code is available at https://github.com/elicassion/StARformer.

(2 min) Stochastic gradient Markov chain Monte Carlo (SGMCMC) is considered the gold standard for Bayesian inference in large-scale models, such as Bayesian neural networks. Since practitioners face speed versus accuracy tradeoffs in these models, variational inference (VI) is often the preferable option. Unfortunately, VI makes strong assumptions on both the factorization and functional form of the posterior. In this work, we propose a new non-parametric variational approximation that makes no assumptions about the approximate posterior's functional form and allows practitioners to specify the exact dependencies the algorithm should respect or break. The approach relies on a new Langevin-type algorithm that operates on a modified energy function, where parts of the latent variables are averaged over samples from earlier iterations of the Markov chain. This way, statistical dependencies can be broken in a controlled way, allowing the chain to mix faster. This scheme can be further modified in a "dropout" manner, leading to even more scalability. By implementing the scheme on a ResNet-20 architecture, we obtain better predictive likelihoods and larger effective sample sizes than full SGMCMC.

(2 min) Data processing is an important step in various natural language processing tasks. As the commonly used datasets in named entity recognition contain only a limited number of samples, it is important to obtain additional labeled data in an efficient and reliable manner. A common practice is to utilize large monolingual unlabeled corpora. Another popular technique is to create synthetic data from the original labeled data (data augmentation). In this work, we investigate the impact of these two methods on the performance of three different named entity recognition tasks.

(0 min) This paper sustains the position that the time has come for thinking of learning machines that conquer visual skills in a truly human-like context, where a few human-like object supervisions are given by vocal interactions and pointing aids only. This likely requires new foundations on computational processes of vision with the final purpose of involving machines in tasks of visual description by living in their own visual environment under simple man-machine linguistic interactions. The challenge consists of developing machines that learn to see without needing to handle visual databases. This might open the doors to a truly orthogonal competitive track concerning deep learning technologies for vision which does not rely on the accumulation of huge visual databases.

(0 min) In contrast to traditional online videos, live multi-streaming supports real-time social interactions between multiple streamers and viewers, such as donations. However, donation and multi-streaming channel recommendations are challenging due to complicated streamer and viewer relations, asymmetric communications, and the tradeoff between personal interests and group interactions. In this paper, we introduce Multi-Stream Party (MSP) and formulate a new multi-streaming recommendation problem, called Donation and MSP Recommendation (DAMRec). We propose Multi-stream Party Recommender System (MARS) to extract latent features via socio-temporal coupled donation-response tensor factorization for donation and MSP recommendations. Experimental results on Twitch and Douyu manifest that MARS significantly outperforms existing recommenders by at least 38.8% in terms of hit ratio and mean average precision.

(0 min) The pretrain-finetune paradigm has shown outstanding performance on many applications of deep learning, where a model is pre-trained on a upstream large dataset (e.g. ImageNet), and is then fine-tuned to different downstream tasks. Though for most cases, the pre-training stage is conducted based on supervised methods, recent works on self-supervised pre-training have shown powerful transferability and even outperform supervised pre-training on multiple downstream tasks. It thus remains an open question how to better generalize supervised pre-training model to downstream tasks. In this paper, we argue that the worse transferability of existing supervised pre-training methods arise from the negligence of valuable intra-class semantic difference. This is because these methods tend to push images from the same class close to each other despite of the large diversity in their visual contents, a problem to which referred as "overfit of upstream tasks". To alleviate this problem, we propose a new supervised pre-training method based on Leave-One-Out K-Nearest-Neighbor, or LOOK for short. It relieves the problem of overfitting upstream tasks by only requiring each image to share its class label with most of its k nearest neighbors, thus allowing each class to exhibit a multi-mode distribution and consequentially preserving part of intra-class difference for better transferring to downstream tasks. We developed efficient implementation of the proposed method that scales well to large datasets. Experimental studies on multiple downstream tasks show that LOOK outperforms other state-of-the-art methods for supervised and self-supervised pre-training.

(0 min) Group equivariant convolutional neural networks (G-CNNs) are generalizations of convolutional neural networks (CNNs) which excel in a wide range of scientific and technical applications by explicitly encoding group symmetries, such as rotations and permutations, in their architectures. Although the success of G-CNNs is driven by the explicit symmetry bias of their convolutional architecture, a recent line of work has proposed that the implicit bias of training algorithms on a particular parameterization (or architecture) is key to understanding generalization for overparameterized neural nets. In this context, we show that $L$-layer full-width linear G-CNNs trained via gradient descent in a binary classification task converge to solutions with low-rank Fourier matrix coefficients, regularized by the $2/L$-Schatten matrix norm. Our work strictly generalizes previous analysis on the implicit bias of linear CNNs to linear G-CNNs over all finite groups, including the challenging setting of non-commutative symmetry groups (such as permutations). We validate our theorems via experiments on a variety of groups and empirically explore more realistic nonlinear networks, which locally capture similar regularization patterns. Finally, we provide intuitive interpretations of our Fourier space implicit regularization results in real space via uncertainty principles.

(0 min) The \textit{lottery ticket hypothesis} (LTH) states that learning on a properly pruned network (the \textit{winning ticket}) improves test accuracy over the original unpruned network. Although LTH has been justified empirically in a broad range of deep neural network (DNN) involved applications like computer vision and natural language processing, the theoretical validation of the improved generalization of a winning ticket remains elusive. To the best of our knowledge, our work, for the first time, characterizes the performance of training a pruned neural network by analyzing the geometric structure of the objective function and the sample complexity to achieve zero generalization error. We show that the convex region near a desirable model with guaranteed generalization enlarges as the neural network model is pruned, indicating the structural importance of a winning ticket. Moreover, when the algorithm for training a pruned neural network is specified as an (accelerated) stochastic gradient descent algorithm, we theoretically show that the number of samples required for achieving zero generalization error is proportional to the number of the non-pruned weights in the hidden layer. With a fixed number of samples, training a pruned neural network enjoys a faster convergence rate to the desired model than training the original unpruned one, providing a formal justification of the improved generalization of the winning ticket. Our theoretical results are acquired from learning a pruned neural network of one hidden layer, while experimental results are further provided to justify the implications in pruning multi-layer neural networks.

(0 min) The integration of algorithmic components into neural architectures has gained increased attention recently, as it allows training neural networks with new forms of supervision such as ordering constraints or silhouettes instead of using ground truth labels. Many approaches in the field focus on the continuous relaxation of a specific task and show promising results in this context. But the focus on single tasks also limits the applicability of the proposed concepts to a narrow range of applications. In this work, we build on those ideas to propose an approach that allows to integrate algorithms into end-to-end trainable neural network architectures based on a general approximation of discrete conditions. To this end, we relax these conditions in control structures such as conditional statements, loops, and indexing, so that resulting algorithms are smoothly differentiable. To obtain meaningful gradients, each relevant variable is perturbed via logistic distributions and the expectation value under this perturbation is approximated. We evaluate the proposed continuous relaxation model on four challenging tasks and show that it can keep up with relaxations specifically designed for each individual task.

(0 min) Heterogeneous information network (HIN) embedding aims to find the representations of nodes that preserve the proximity between entities of different nature. A family of approaches that are wildly adopted applies random walk to generate a sequence of heterogeneous context, from which the embedding is learned. However, due to the multipartite graph structure of HIN, hub nodes tend to be over-represented in the sampled sequence, giving rise to imbalanced samples of the network. Here we propose a new embedding method CoarSAS2hvec. The self-avoid short sequence sampling with the HIN coarsening procedure (CoarSAS) is utilized to better collect the rich information in HIN. An optimized loss function is used to improve the performance of the HIN structure embedding. CoarSAS2hvec outperforms nine other methods in two different tasks on four real-world data sets. The ablation study confirms that the samples collected by CoarSAS contain richer information of the network compared with those by other methods, which is characterized by a higher information entropy. Hence, the traditional loss function applied to samples by CoarSAS can also yield improved results. Our work addresses a limitation of the random-walk-based HIN embedding that has not been emphasized before, which can shed light on a range of problems in HIN analyses.

(0 min) We investigate a general matrix factorization for deviance-based losses, extending the ubiquitous singular value decomposition beyond squared error loss. While similar approaches have been explored before, here we propose an efficient algorithm that is flexible enough to allow for structural zeros and entry weights. Moreover, we provide theoretical support for these decompositions by (i) showing strong consistency under a generalized linear model setup, (ii) checking the adequacy of a chosen exponential family via a generalized Hosmer-Lemeshow test, and (iii) determining the rank of the decomposition via a maximum eigenvalue gap method. To further support our findings, we conduct simulation studies to assess robustness to decomposition assumptions and extensive case studies using benchmark datasets from image face recognition, natural language processing, network analysis, and biomedical studies. Our theoretical and empirical results indicate that the proposed decomposition is more flexible, general, and can provide improved performance when compared to traditional methods.

(0 min) Latent variable discovery is a central problem in data analysis with a broad range of applications in applied science. In this work, we consider data given as an invertible mixture of two statistically independent components, and assume that one of the components is observed while the other is hidden. Our goal is to recover the hidden component. For this purpose, we propose an autoencoder equipped with a discriminator. Unlike the standard nonlinear ICA problem, which was shown to be non-identifiable, in the special case of ICA we consider here, we show that our approach can recover the component of interest up to entropy-preserving transformation. We demonstrate the performance of the proposed approach on several datasets, including image synthesis, voice cloning, and fetal ECG extraction.

(0 min) Given the popularity of social media and the notion of it being a platform encouraging free speech, it has become an open playground for user (bot) accounts trying to manipulate other users using these platforms. Social bots not only learn human conversations, manners, and presence but also manipulate public opinion, act as scammers, manipulate stock markets, etc. There has been evidence of bots manipulating the election results which can be a great threat to the whole nation and hence the whole world. So identification and prevention of such campaigns that release or create the bots have become critical to tackling it at its source of origin. Our goal is to leverage semantic web mining techniques to identify fake bots or accounts involved in these activities.

(0 min) Many software engineering tasks, such as testing, and anomaly detection can benefit from the ability to infer a behavioral model of the software.Most existing inference approaches assume access to code to collect execution sequences. In this paper, we investigate a black-box scenario, where the system under analysis cannot be instrumented, in this granular fashion.This scenario is particularly prevalent with control systems' log analysis in the form of continuous signals. In this situation, an execution trace amounts to a multivariate time-series of input and output signals, where different states of the system correspond to different `phases` in the time-series. The main challenge is to detect when these phase changes take place. Unfortunately, most existing solutions are either univariate, make assumptions on the data distribution, or have limited learning power.Therefore, we propose a hybrid deep neural network that accepts as input a multivariate time series and applies a set of convolutional and recurrent layers to learn the non-linear correlations between signals and the patterns over time.We show how this approach can be used to accurately detect state changes, and how the inferred models can be successfully applied to transfer-learning scenarios, to accurately process traces from different products with similar execution characteristics. Our experimental results on two UAV autopilot case studies indicate that our approach is highly accurate (over 90% F1 score for state classification) and significantly improves baselines (by up to 102% for change point detection).Using transfer learning we also show that up to 90% of the maximum achievable F1 scores in the open-source case study can be achieved by reusing the trained models from the industrial case and only fine tuning them using as low as 5 labeled samples, which reduces the manual labeling effort by 98%.

(0 min) SARS-CoV-2, like any other virus, continues to mutate as it spreads, according to an evolutionary process. Unlike any other virus, the number of currently available sequences of SARS-CoV-2 in public databases such as GISAID is already several million. This amount of data has the potential to uncover the evolutionary dynamics of a virus like never before. However, a million is already several orders of magnitude beyond what can be processed by the traditional methods designed to reconstruct a virus's evolutionary history, such as those that build a phylogenetic tree. Hence, new and scalable methods will need to be devised in order to make use of the ever increasing number of viral sequences being collected. Since identifying variants is an important part of understanding the evolution of a virus, in this paper, we propose an approach based on clustering sequences to identify the current major SARS-CoV-2 variants. Using a $k$-mer based feature vector generation and efficient feature selection methods, our approach is effective in identifying variants, as well as being efficient and scalable to millions of sequences. Such a clustering method allows us to show the relative proportion of each variant over time, giving the rate of spread of each variant in different locations -- something which is important for vaccine development and distribution. We also compute the importance of each amino acid position of the spike protein in identifying a given variant in terms of information gain. Positions of high variant-specific importance tend to agree with those reported by the USA's Centers for Disease Control and Prevention (CDC), further demonstrating our approach.

(0 min) The ability to form complex plans based on raw visual input is a litmus test for current capabilities of artificial intelligence, as it requires a seamless combination of visual processing and abstract algorithmic execution, two traditionally separate areas of computer science. A recent surge of interest in this field brought advances that yield good performance in tasks ranging from arcade games to continuous control; these methods however do not come without significant issues, such as limited generalization capabilities and difficulties when dealing with combinatorially hard planning instances. Our contribution is two-fold: (i) we present a method that learns to represent its environment as a latent graph and leverages state reidentification to reduce the complexity of finding a good policy from exponential to linear (ii) we introduce a set of lightweight environments with an underlying discrete combinatorial structure in which planning is challenging even for humans. Moreover, we show that our methods achieves strong empirical generalization to variations in the environment, even across highly disadvantaged regimes, such as "one-shot" planning, or in an offline RL paradigm which only provides low-quality trajectories.

(0 min) While theoretically appealing, the application of the Wasserstein distance to large-scale machine learning problems has been hampered by its prohibitive computational cost. The sliced Wasserstein distance and its variants improve the computational efficiency through the random projection, yet they suffer from low accuracy if the number of projections is not sufficiently large, because the majority of projections result in trivially small values. In this work, we propose a new family of distance metrics, called augmented sliced Wasserstein distances (ASWDs), constructed by first mapping samples to higher-dimensional hypersurfaces parameterized by neural networks. It is derived from a key observation that (random) linear projections of samples residing on these hypersurfaces would translate to much more flexible nonlinear projections in the original sample space, so they can capture complex structures of the data distribution. We show that the hypersurfaces can be optimized by gradient ascent efficiently. We provide the condition under which the ASWD is a valid metric and show that this can be obtained by an injective neural network architecture. Numerical results demonstrate that the ASWD significantly outperforms other Wasserstein variants for both synthetic and real-world problems.

(0 min) Communication between embodied AI agents has received increasing attention in recent years. Despite its use, it is still unclear whether the learned communication is interpretable and grounded in perception. To study the grounding of emergent forms of communication, we first introduce the collaborative multi-object navigation task CoMON. In this task, an oracle agent has detailed environment information in the form of a map. It communicates with a navigator agent that perceives the environment visually and is tasked to find a sequence of goals. To succeed at the task, effective communication is essential. CoMON hence serves as a basis to study different communication mechanisms between heterogeneous agents, that is, agents with different capabilities and roles. We study two common communication mechanisms and analyze their communication patterns through an egocentric and spatial lens. We show that the emergent communication can be grounded to the agent observations and the spatial structure of the 3D environment. Video summary: https://youtu.be/kLv2rxO9t0g

(0 min) Approximate linear programs (ALPs) are well-known models based on value function approximations (VFAs) to obtain policies and lower bounds on the optimal policy cost of discounted-cost Markov decision processes (MDPs). Formulating an ALP requires (i) basis functions, the linear combination of which defines the VFA, and (ii) a state-relevance distribution, which determines the relative importance of different states in the ALP objective for the purpose of minimizing VFA error. Both these choices are typically heuristic: basis function selection relies on domain knowledge while the state-relevance distribution is specified using the frequency of states visited by a heuristic policy. We propose a self-guided sequence of ALPs that embeds random basis functions obtained via inexpensive sampling and uses the known VFA from the previous iteration to guide VFA computation in the current iteration. Self-guided ALPs mitigate the need for domain knowledge during basis function selection as well as the impact of the initial choice of the state-relevance distribution, thus significantly reducing the ALP implementation burden. We establish high probability error bounds on the VFAs from this sequence and show that a worst-case measure of policy performance is improved. We find that these favorable implementation and theoretical properties translate to encouraging numerical results on perishable inventory control and options pricing applications, where self-guided ALP policies improve upon policies from problem-specific methods. More broadly, our research takes a meaningful step toward application-agnostic policies and bounds for MDPs.

(0 min) Vision-and-language navigation (VLN) is a trending topic which aims to navigate an intelligent agent to an expected position through natural language instructions. This work addresses the task of VLN from a previously-ignored aspect, namely the spatial route prior of the navigation scenes. A critically enabling innovation of this work is explicitly considering the spatial route prior under several different VLN settings. In a most information-rich case of knowing environment maps and admitting shortest-path prior, we observe that given an origin-destination node pair, the internal route can be uniquely determined. Thus, VLN can be effectively formulated as an ordinary classification problem over all possible destination nodes in the scenes. Furthermore, we relax it to other more general VLN settings, proposing a sequential-decision variant (by abandoning the shortest-path route prior) and an explore-and-exploit scheme (for addressing the case of not knowing the environment maps) that curates a compact and informative sub-graph to exploit. As reported by [34], the performance of VLN methods has been stuck at a plateau in past two years. Even with increased model complexity, the state-of-the-art success rate on R2R validation-unseen set has stayed around 62% for single-run and 73% for beam-search with model-ensemble. We have conducted comprehensive evaluations on both R2R and R4R, and surprisingly found that utilizing the spatial route priors may be the key of breaking above-mentioned performance ceiling. For example, on R2R validation-unseen set, when the number of discrete nodes explored is about 40, our single-model success rate reaches 73%, and increases to 78% if a Speaker model is ensembled, which significantly outstrips previous state-of-the-art VLN-BERT with 3 models ensembled.

(0 min) Performance of recommender systems (RS) relies heavily on the amount of training data available. This poses a chicken-and-egg problem for early-stage products, whose amount of data, in turn, relies on the performance of their RS. On the other hand, zero-shot learning promises some degree of generalization from an old dataset to an entirely new dataset. In this paper, we explore the possibility of zero-shot learning in RS. We develop an algorithm, dubbed ZEro-Shot Recommenders (ZESRec), that is trained on an old dataset and generalize to a new one where there are neither overlapping users nor overlapping items, a setting that contrasts typical cross-domain RS that has either overlapping users or items. Different from categorical item indices, i.e., item ID, in previous methods, ZESRec uses items' natural-language descriptions (or description embeddings) as their continuous indices, and therefore naturally generalize to any unseen items. In terms of users, ZESRec builds upon recent advances on sequential RS to represent users using their interactions with items, thereby generalizing to unseen users as well. We study three pairs of real-world RS datasets and demonstrate that ZESRec can successfully enable recommendations in such a zero-shot setting, opening up new opportunities for resolving the chicken-and-egg problem for data-scarce startups or early-stage products.

(0 min) In this paper we propose a novel method to forecast the result of elections using only official results of previous ones. It is based on the voter model with stubborn nodes and uses theoretical results developed in a previous work of ours. We look at popular vote shares for the Conservative and Labour parties in the UK and the Republican and Democrat parties in the US. We are able to perform time-evolving estimates of the model parameters and use these to forecast the vote shares for each party in any election. We obtain a mean absolute error of 4.74\%. As a side product, our parameters estimates provide meaningful insight on the political landscape, informing us on the proportion of voters that are strong supporters of each of the considered parties.

(0 min) Recognising previously visited locations is an important, but unsolved, task in autonomous navigation. Current visual place recognition (VPR) benchmarks typically challenge models to recover the position of a query image (or images) from sequential datasets that include both spatial and temporal components. Recently, Echo State Network (ESN) varieties have proven particularly powerful at solving machine learning tasks that require spatio-temporal modelling. These networks are simple, yet powerful neural architectures that -- exhibiting memory over multiple time-scales and non-linear high-dimensional representations -- can discover temporal relations in the data while still maintaining linearity in the learning. In this paper, we present a series of ESNs and analyse their applicability to the VPR problem. We report that the addition of ESNs to pre-processed convolutional neural networks led to a dramatic boost in performance in comparison to non-recurrent networks in four standard benchmarks (GardensPoint, SPEDTest, ESSEX3IN1, Nordland) demonstrating that ESNs are able to capture the temporal structure inherent in VPR problems. Moreover, we show that ESNs can outperform class-leading VPR models which also exploit the sequential dynamics of the data. Finally, our results demonstrate that ESNs also improve generalisation abilities, robustness, and accuracy further supporting their suitability to VPR applications.

(0 min) A recommender system learns to predict the user-specific preference or intention over many items simultaneously for all users, making personalized recommendations based on a relatively small number of observations. One central issue is how to leverage three-way interactions, referred to as user-item-stage dependencies on a monotonic chain of events, to enhance the prediction accuracy. A monotonic chain of events occurs, for instance, in an article sharing dataset, where a ``follow'' action implies a ``like'' action, which in turn implies a ``view'' action. In this article, we develop a multistage recommender system utilizing a two-level monotonic property characterizing a monotonic chain of events for personalized prediction. Particularly, we derive a large-margin classifier based on a nonnegative additive latent factor model in the presence of a high percentage of missing observations, particularly between stages, reducing the number of model parameters for personalized prediction while guaranteeing prediction consistency. On this ground, we derive a regularized cost function to learn user-specific behaviors at different stages, linking decision functions to numerical and categorical covariates to model user-item-stage interactions. Computationally, we derive an algorithm based on blockwise coordinate descent. Theoretically, we show that the two-level monotonic property enhances the accuracy of learning as compared to a standard method treating each stage individually and an ordinal method utilizing only one-level monotonicity. Finally, the proposed method compares favorably with existing methods in simulations and an article sharing dataset.

(0 min) Over the past decade, tremendous progress has been made in inventing new RecSys methods. However, one of the fundamental problems of the RecSys research community remains the lack of applied datasets and benchmarks with well-defined evaluation rules and metrics to test these novel approaches. In this article, we present the TTRS - Tinkoff Transactions Recommender System benchmark. This financial transaction benchmark contains over 2 million interactions between almost 10,000 users and more than 1,000 merchant brands over 14 months. To the best of our knowledge, this is the first publicly available financial transactions dataset. To make it more suitable for possible applications, we provide a complete description of the data collection pipeline, its preprocessing, and the resulting dataset statistics. We also present a comprehensive comparison of the current popular RecSys methods on the next-period recommendation task and conduct a detailed analysis of their performance against various metrics and recommendation goals. Last but not least, we also introduce Personalized Item-Frequencies-based Model (Re)Ranker - PIFMR, a simple yet powerful approach that has proven to be the most effective for the benchmarked tasks.

(0 min) During the past decade, many anomaly detection approaches have been introduced in different fields such as network monitoring, fraud detection, and intrusion detection. However, they require understanding of data pattern and often need a long off-line period to build a model or network for the target data. Providing real-time and proactive anomaly detection for streaming time series without human intervention and domain knowledge is highly valuable since it greatly reduces human effort and enables appropriate countermeasures to be undertaken before a disastrous damage, failure, or other harmful event occurs. However, this issue has not been well studied yet. To address it, this paper proposes RePAD, which is a Real-time Proactive Anomaly Detection algorithm for streaming time series based on Long Short-Term Memory (LSTM). RePAD utilizes short-term historic data points to predict and determine whether or not the upcoming data point is a sign that an anomaly is likely to happen in the near future. By dynamically adjusting the detection threshold over time, RePAD is able to tolerate minor pattern change in time series and detect anomalies either proactively or on time. Experiments based on two time series datasets collected from the Numenta Anomaly Benchmark demonstrate that RePAD is able to proactively detect anomalies and provide early warnings in real time without human intervention and domain knowledge.

(0 min) Many real-world applications of multi-agent reinforcement learning (RL), such as multi-robot navigation and decentralized control of cyber-physical systems, involve the cooperation of agents as a team with aligned objectives. We study multi-agent RL in the most basic cooperative setting -- Markov teams -- a class of Markov games where the cooperating agents share a common reward. We propose an algorithm in which each agent independently runs stage-based V-learning (a Q-learning style algorithm) to efficiently explore the unknown environment, while using a stochastic gradient descent (SGD) subroutine for policy updates. We show that the agents can learn an $\epsilon$-approximate Nash equilibrium policy in at most $\propto\widetilde{O}(1/\epsilon^4)$ episodes. Our results advocate the use of a novel \emph{stage-based} V-learning approach to create a stage-wise stationary environment. We also show that under certain smoothness assumptions of the team, our algorithm can achieve a nearly \emph{team-optimal} Nash equilibrium. Simulation results corroborate our theoretical findings. One key feature of our algorithm is being \emph{decentralized}, in the sense that each agent has access to only the state and its local actions, and is even \emph{oblivious} to the presence of the other agents. Neither communication among teammates nor coordination by a central controller is required during learning. Hence, our algorithm can readily generalize to an arbitrary number of agents, without suffering from the exponential dependence on the number of agents.

(0 min) In knowledge graph embedding, the theoretical relationship between the softmax cross-entropy and negative sampling loss functions has not been investigated. This makes it difficult to fairly compare the results of the two different loss functions. We attempted to solve this problem by using the Bregman divergence to provide a unified interpretation of the softmax cross-entropy and negative sampling loss functions. Under this interpretation, we can derive theoretical findings for fair comparison. Experimental results on the FB15k-237 and WN18RR datasets show that the theoretical findings are valid in practical settings.

(0 min) We consider the task of solving generic inverse problems, where one wishes to determine the hidden parameters of a natural system that will give rise to a particular set of measurements. Recently many new approaches based upon deep learning have arisen generating impressive results. We conceptualize these models as different schemes for efficiently, but randomly, exploring the space of possible inverse solutions. As a result, the accuracy of each approach should be evaluated as a function of time rather than a single estimated solution, as is often done now. Using this metric, we compare several state-of-the-art inverse modeling approaches on four benchmark tasks: two existing tasks, one simple task for visualization and one new task from metamaterial design. Finally, inspired by our conception of the inverse problem, we explore a solution that uses a deep learning model to approximate the forward model, and then uses backpropagation to search for good inverse solutions. This approach, termed the neural-adjoint, achieves the best performance in many scenarios.

(0 min) Many popular learning-rate schedules for deep neural networks combine a decaying trend with local perturbations that attempt to escape saddle points and bad local minima. We derive convergence guarantees for bandwidth-based step-sizes, a general class of learning rates that are allowed to vary in a banded region. This framework includes many popular cyclic and non-monotonic step-sizes for which no theoretical guarantees were previously known. We provide worst-case guarantees for SGD on smooth non-convex problems under several bandwidth-based step sizes, including stagewise $1/\sqrt{t}$ and the popular step-decay (constant and then drop by a constant), which is also shown to be optimal. Moreover, we show that its momentum variant converges as fast as SGD with the bandwidth-based step-decay step-size. Finally, we propose novel step-size schemes in the bandwidth-based family and verify their efficiency on several deep neural network training tasks.

(0 min) Knowledge embedded in the weights of the artificial neural network can be used to improve the network structure, such as in network compression. However, the knowledge is set up by hand, which may not be very accurate, and relevant information may be overlooked. Inspired by how learning works in the mammalian brain, we mine the knowledge contained in the weights of the neural network toward automatic architecture learning in this paper. We introduce a switcher neural network (SNN) that uses as inputs the weights of a task-specific neural network (called TNN for short). By mining the knowledge contained in the weights, the SNN outputs scaling factors for turning off and weighting neurons in the TNN. To optimize the structure and the parameters of TNN simultaneously, the SNN and TNN are learned alternately under the same performance evaluation of TNN using stochastic gradient descent. We test our method on widely used datasets and popular networks in classification applications. In terms of accuracy, we outperform baseline networks and other structure learning methods stably and significantly. At the same time, we compress the baseline networks without introducing any sparse induction mechanism, and our method, in particular, leads to a lower compression rate when dealing with simpler baselines or more difficult tasks. These results demonstrate that our method can produce a more reasonable structure.

(0 min) Bayesian learning via Stochastic Gradient Langevin Dynamics (SGLD) has been suggested for differentially private learning. While previous research provides differential privacy bounds for SGLD when close to convergence or at the initial steps of the algorithm, the question of what differential privacy guarantees can be made in between remains unanswered. This interim region is essential, especially for Bayesian neural networks, as it is hard to guarantee convergence to the posterior. This paper will show that using SGLD might result in unbounded privacy loss for this interim region, even when sampling from the posterior is as differentially private as desired.

(0 min) Real world tournaments are almost always intransitive. Recent works have noted that parametric models which assume $d$ dimensional node representations can effectively model intransitive tournaments. However, nothing is known about the structure of the class of tournaments that arise out of any fixed $d$ dimensional representations. In this work, we develop a novel theory for understanding parametric tournament representations. Our first contribution is to structurally characterize the class of tournaments that arise out of $d$ dimensional representations. We do this by showing that these tournament classes have forbidden configurations which must necessarily be union of flip classes, a novel way to partition the set of all tournaments. We further characterise rank $2$ tournaments completely by showing that the associated forbidden flip class contains just $2$ tournaments. Specifically, we show that the rank $2$ tournaments are equivalent to locally-transitive tournaments. This insight allows us to show that the minimum feedback arc set problem on this tournament class can be solved using the standard Quicksort procedure. For a general rank $d$ tournament class, we show that the flip class associated with a coned-doubly regular tournament of size $\mathcal{O}(\sqrt{d})$ must be a forbidden configuration. To answer a dual question, using a celebrated result of \cite{forster}, we show a lower bound of $\mathcal{O}(\sqrt{n})$ on the minimum dimension needed to represent all tournaments on $n$ nodes. For any given tournament, we show a novel upper bound on the smallest representation dimension that depends on the least size of the number of unique nodes in any feedback arc set of the flip class associated with a tournament. We show how our results also shed light on upper bound of sign-rank of matrices.

(0 min) Anomaly detection is the process of identifying unexpected events or ab-normalities in data, and it has been applied in many different areas such as system monitoring, fraud detection, healthcare, intrusion detection, etc. Providing real-time, lightweight, and proactive anomaly detection for time series with neither human intervention nor domain knowledge could be highly valuable since it reduces human effort and enables appropriate countermeasures to be undertaken before a disastrous event occurs. To our knowledge, RePAD (Real-time Proactive Anomaly Detection algorithm) is a generic approach with all above-mentioned features. To achieve real-time and lightweight detection, RePAD utilizes Long Short-Term Memory (LSTM) to detect whether or not each upcoming data point is anomalous based on short-term historical data points. However, it is unclear that how different amounts of historical data points affect the performance of RePAD. Therefore, in this paper, we investigate the impact of different amounts of historical data on RePAD by introducing a set of performance metrics that cover novel detection accuracy measures, time efficiency, readiness, and resource consumption, etc. Empirical experiments based on real-world time series datasets are conducted to evaluate RePAD in different scenarios, and the experimental results are presented and discussed.

(0 min) A major challenge in structured prediction is to represent the interdependencies within output structures. When outputs are structured as sequences, linear-chain conditional random fields (CRFs) are a widely used model class which can learn \textit{local} dependencies in the output. However, the CRF's Markov assumption makes it impossible for CRFs to represent distributions with \textit{nonlocal} dependencies, and standard CRFs are unable to respect nonlocal constraints of the data (such as global arity constraints on output labels). We present a generalization of CRFs that can enforce a broad class of constraints, including nonlocal ones, by specifying the space of possible output structures as a regular language $\mathcal{L}$. The resulting regular-constrained CRF (RegCCRF) has the same formal properties as a standard CRF, but assigns zero probability to all label sequences not in $\mathcal{L}$. Notably, RegCCRFs can incorporate their constraints during training, while related models only enforce constraints during decoding. We prove that constrained training is never worse than constrained decoding, and show empirically that it can be substantially better in practice. Additionally, we demonstrate a practical benefit on downstream tasks by incorporating a RegCCRF into a deep neural model for semantic role labeling, exceeding state-of-the-art results on a standard dataset.

(0 min) Detecting the spread of coronavirus will go a long way toward reducing human and economic loss. Unfortunately, existing Epidemiological models used for COVID 19 prediction models are too slow and fail to capture the COVID-19 development in detail. This research uses Partial Differential Equations to improve the processing speed and accuracy of forecasting of COVID 19 governed by SEIRD model equations. The dynamics of COVID 19 were extracted using Convolutional Neural Networks and Deep Residual Recurrent Neural Networks from data simulated using PDEs. The DRRNNs accuracy is measured using Mean Squared Error. The DRRNNs COVID-19 prediction model has been shown to have accurate COVID-19 predictions. In addition, we concluded that DR-RNNs can significantly advance the ability to support decision-making in real time COVID-19 prediction.

(0 min) Graph Convolutional Networks (GCNs) are a popular method from graph representation learning that have proved effective for tasks like node classification tasks. Although typical GCN models focus on classifying nodes within a static graph, several recent variants propose node classification in dynamic graphs whose topologies and node attributes change over time, e.g., social networks with dynamic relationships, or literature citation networks with changing co-authorships. These works, however, do not fully address the challenge of flexibly assigning different importance to snapshots of the graph at different times, which depending on the graph dynamics may have more or less predictive power on the labels. We address this challenge by proposing a new method, GCN-SE, that attaches a set of learnable attention weights to graph snapshots at different times, inspired by Squeeze and Excitation Net (SE-Net). We show that GCN-SE outperforms previously proposed node classification methods on a variety of graph datasets. To verify the effectiveness of the attention weight in determining the importance of different graph snapshots, we adapt perturbation-based methods from the field of explainable machine learning to graphical settings and evaluate the correlation between the attention weights learned by GCN-SE and the importance of different snapshots over time. These experiments demonstrate that GCN-SE can in fact identify different snapshots' predictive power for dynamic node classification.

(0 min) This work demonstrates that using the objective with independence assumption for modelling the span probability $P(a_s,a_e) = P(a_s)P(a_e)$ of span starting at position $a_s$ and ending at position $a_e$ has adverse effects. Therefore we propose multiple approaches to modelling joint probability $P(a_s,a_e)$ directly. Among those, we propose a compound objective, composed from the joint probability while still keeping the objective with independence assumption as an auxiliary objective. We find that the compound objective is consistently superior or equal to other assumptions in exact match. Additionally, we identified common errors caused by the assumption of independence and manually checked the counterpart predictions, demonstrating the impact of the compound objective on the real examples. Our findings are supported via experiments with three extractive QA models (BIDAF, BERT, ALBERT) over six datasets and our code, individual results and manual analysis are available online.

(0 min) Twitter is a social network that offers a rich and interesting source of information challenging to retrieve and analyze. Twitter data can be accessed using a REST API. The available operations allow retrieving tweets on the basis of a set of keywords but with limitations such as the number of calls per minute and the size of results. Besides, there is no control on retrieved results and finding tweets which are relevant to a specific topic is a big issue. Given these limitations, it is important that the query keywords cover unambiguously the topic of interest in order to both reach the relevant answers and decrease the number of API calls. In this paper, we introduce a new crawling algorithm called "SmartTwitter Crawling" (STiC) that retrieves a set of tweets related to a target topic. In this algorithm, we take an initial keyword query and enrich it using a set of additional keywords that come from different data sources. STiC algorithm relies on a DFS search in Twittergraph where each reached tweet is considered if it is relevant with the query keywords using a scoring, updated throughout the whole crawling process. This scoring takes into account the tweet text, hashtags and the users who have posted the tweet, replied to the tweet, been mentioned in the tweet or retweeted the tweet. Given this score, STiC is able to select relevant tweets in each iteration and continue by adding the related valuable tweets. Several experiments have been achieved for different kinds of queries, the results showedthat the precision increases compared to a simple BFS search.

(0 min) The Internet of Things (IoT) is reshaping modern society by allowing a decent number of RF devices to connect and share information through RF channels. However, such an open nature also brings obstacles to surveillance. For alleviation, a surveillance oracle, or a cognitive communication entity needs to identify and confirm the appearance of known or unknown signal sources in real-time. In this paper, we provide a deep learning framework for RF signal surveillance. Specifically, we jointly integrate the Deep Neural Networks (DNNs) and Quickest Detection (QD) to form a sequential signal surveillance scheme. We first analyze the latent space characteristic of neural network classification models, and then we leverage the response characteristics of DNN classifiers and propose a novel method to transform existing DNN classifiers into performance-assured binary abnormality detectors. In this way, we seamlessly integrate the DNNs with the parametric quickest detection. Finally, we propose an enhanced Elastic Weight Consolidation (EWC) algorithm with better numerical stability for DNNs in signal surveillance systems to evolve incrementally, we demonstrate that the zero-bias DNN is superior to regular DNN models considering incremental learning and decision fairness. We evaluated the proposed framework using real signal datasets and we believe this framework is helpful in developing a trustworthy IoT ecosystem.

(0 min) Obtaining standardized crowdsourced benchmark of computational methods is a major issue in scientific communities. Dedicated frameworks enabling fair continuous benchmarking in a unified environment are yet to be developed. Here we introduce Codabench, an open-sourced, community-driven platform for benchmarking algorithms or software agents versus datasets or tasks. A public instance of Codabench is open to everyone, free of charge, and allows benchmark organizers to compare fairly submissions, under the same setting (software, hardware, data, algorithms), with custom protocols and data formats. Codabench has unique features facilitating the organization of benchmarks flexibly, easily and reproducibly. Firstly, it supports code submission and data submission for testing on dedicated compute workers, which can be supplied by the benchmark organizers. This makes the system scalable, at low cost for the platform providers. Secondly, Codabench benchmarks are created from self-contained bundles, which are zip files containing a full description of the benchmark in a configuration file (following a well-defined schema), documentation pages, data, ingestion and scoring programs, making benchmarks reusable and portable. The Codabench documentation includes many examples of bundles that can serve as templates. Thirdly, Codabench uses dockers for each task's running environment to make results reproducible. Codabench has been used internally and externally with more than 10 applications during the past 6 months. As illustrative use cases, we introduce 4 diverse benchmarks covering Graph Machine Learning, Cancer Heterogeneity, Clinical Diagnosis and Reinforcement Learning.

(0 min) Stereoselective reactions (both chemical and enzymatic reactions) have been essential for origin of life, evolution, human biology and medicine. Since late 1960s, there have been numerous successes in the exciting new frontier of asymmetric catalysis. However, most industrial and academic asymmetric catalysis nowadays do follow the trial-and-error model, since the energetic difference for success or failure in asymmetric catalysis is incredibly small. Our current understanding about stereoselective reactions is mostly qualitative that stereoselectivity arises from differences in steric effects and electronic effects in multiple competing mechanistic pathways. Quantitatively understanding and modulating the stereoselectivity of for a given chemical reaction still remains extremely difficult. As a proof of principle, we herein present a novel machine learning technique, which combines a LASSO model and two Random Forest model via two Gaussian Mixture models, for quantitatively predicting stereoselectivity of chemical reactions. Compared to the recent ground-breaking approach [1], our approach is able to capture interactions between features and exploit complex data distributions, which are important for predicting stereoselectivity. Experimental results on a recently published dataset demonstrate that our approach significantly outperform [1]. The insight obtained from our results provide a solid foundation for further exploration of other synthetically valuable yet mechanistically intriguing stereoselective reactions.

(0 min) In the study of collective motion, it is common practice to collect movement information at the level of the group to infer the characteristics of the individual agents and their interactions. However, it is not clear whether one can always correctly infer individual characteristics from movement data of the collective. We investigate this question in the context of a composite crowd with two groups of agents, each with its own desired direction of motion. A simple observer attempts to classify an agent into its group based on its movement information. However, collective effects such as collisions, entrainment of agents, formation of lanes and clusters, etc. render the classification problem non-trivial, and lead to misclassifications. Based on our understanding of these effects, we propose a new observer algorithm that infers, based only on observed movement information, how the local neighborhood aids or hinders agent movement. Unlike a traditional supervised learning approach, this algorithm is based on physical insights and scaling arguments, and does not rely on training-data. This new observer improves classification performance and is able to differentiate agents belonging to different groups even when their motion is identical. Data-agnostic approaches like this have relevance to a large class of real-world problems where clean, labeled data is difficult to obtain, and is a step towards hybrid approaches that integrate both data and domain knowledge.

(0 min) Natural language understanding tasks such as open-domain question answering often require retrieving and assimilating factual information from multiple sources. We propose to address this problem by integrating a semi-parametric representation of a large text corpus into a Transformer model as a source of factual knowledge. Specifically, our method represents knowledge with `mention memory', a table of dense vector representations of every entity mention in a corpus. The proposed model - TOME - is a Transformer that accesses the information through internal memory layers in which each entity mention in the input passage attends to the mention memory. This approach enables synthesis of and reasoning over many disparate sources of information within a single Transformer model. In experiments using a memory of 150 million Wikipedia mentions, TOME achieves strong performance on several open-domain knowledge-intensive tasks, including the claim verification benchmarks HoVer and FEVER and several entity-based QA benchmarks. We also show that the model learns to attend to informative mentions without any direct supervision. Finally we demonstrate that the model can generalize to new unseen entities by updating the memory without retraining.

(0 min) Representation learning on temporal interaction graphs (TIG) is to model complex networks with the dynamic evolution of interactions arising in a broad spectrum of problems. Existing dynamic embedding methods on TIG discretely update node embeddings merely when an interaction occurs. They fail to capture the continuous dynamic evolution of embedding trajectories of nodes. In this paper, we propose a two-module framework named ConTIG, a continuous representation method that captures the continuous dynamic evolution of node embedding trajectories. With two essential modules, our model exploit three-fold factors in dynamic networks which include latest interaction, neighbor features and inherent characteristics. In the first update module, we employ a continuous inference block to learn the nodes' state trajectories by learning from time-adjacent interaction patterns between node pairs using ordinary differential equations. In the second transform module, we introduce a self-attention mechanism to predict future node embeddings by aggregating historical temporal interaction information. Experiments results demonstrate the superiority of ConTIG on temporal link prediction, temporal node recommendation and dynamic node classification tasks compared with a range of state-of-the-art baselines, especially for long-interval interactions prediction.

(0 min) The goal of this paper is to develop a practical and general-purpose approach to construct confidence intervals for differentially private parametric estimation. We find that the parametric bootstrap is a simple and effective solution. It cleanly reasons about variability of both the data sample and the randomized privacy mechanism and applies "out of the box" to a wide class of private estimation routines. It can also help correct bias caused by clipping data to limit sensitivity. We prove that the parametric bootstrap gives consistent confidence intervals in two broadly relevant settings, including a novel adaptation to linear regression that avoids accessing the covariate data multiple times. We demonstrate its effectiveness for a variety of estimators, and find that it provides confidence intervals with good coverage even at modest sample sizes and performs better than alternative approaches.

(0 min) Optimizing traffic signal control (TSC) at intersections continues to pose a challenging problem, particularly for large-scale traffic networks. It has been shown in past research that it is feasible to optimize the operations of individual TSC systems or a small number of such systems. However, it has been computationally difficult to scale these solution approaches to large networks partly due to the curse of dimensionality that is encountered as the number of intersections increases. Fortunately, recent studies have recognized the potential of exploiting advancements in deep and reinforcement learning to address this problem, and some preliminary successes have been achieved in this regard. However, facilitating such intelligent solution approaches may require large amounts of infrastructural investments such as roadside units (RSUs) and drones in order to ensure thorough connectivity across all intersections in large networks, an investment that may be burdensome for agencies to undertake. As such, this study builds on recent work to present a scalable TSC model that may reduce the number of required enabling infrastructure. This is achieved using graph attention networks (GATs) to serve as the neural network for deep reinforcement learning, which aids in maintaining the graph topology of the traffic network while disregarding any irrelevant or unnecessary information. A case study is carried out to demonstrate the effectiveness of the proposed model, and the results show much promise. The overall research outcome suggests that by decomposing large networks using fog-nodes, the proposed fog-based graphic RL (FG-RL) model can be easily applied to scale into larger traffic networks.

(0 min) This paper addresses the problem of learning an equilibrium efficiently in general-sum Markov games through decentralized multi-agent reinforcement learning. Given the fundamental difficulty of calculating a Nash equilibrium (NE), we instead aim at finding a coarse correlated equilibrium (CCE), a solution concept that generalizes NE by allowing possible correlations among the agents' strategies. We propose an algorithm in which each agent independently runs optimistic V-learning (a variant of Q-learning) to efficiently explore the unknown environment, while using a stabilized online mirror descent (OMD) subroutine for policy updates. We show that the agents can find an $\epsilon$-approximate CCE in at most $\widetilde{O}( H^6S A /\epsilon^2)$ episodes, where $S$ is the number of states, $A$ is the size of the largest individual action space, and $H$ is the length of an episode. This appears to be the first sample complexity result for learning in generic general-sum Markov games. Our results rely on a novel investigation of an anytime high-probability regret bound for OMD with a dynamic learning rate and weighted regret, which would be of independent interest. One key feature of our algorithm is that it is fully \emph{decentralized}, in the sense that each agent has access to only its local information, and is completely oblivious to the presence of others. This way, our algorithm can readily scale up to an arbitrary number of agents, without suffering from the exponential dependence on the number of agents.

(0 min) In nearest-neighbor classification problems, a set of $d$-dimensional training points are given, each with a known classification, and are used to infer unknown classifications of other points by using the same classification as the nearest training point. A training point is relevant if its omission from the training set would change the outcome of some of these inferences. We provide a simple algorithm for thinning a training set down to its subset of relevant points, using as subroutines algorithms for finding the minimum spanning tree of a set of points and for finding the extreme points (convex hull vertices) of a set of points. The time bounds for our algorithm, in any constant dimension $d\ge 3$, improve on a previous algorithm for the same problem by Clarkson (FOCS 1994).

(0 min) Robots of the future are going to exhibit increasingly human-like and super-human intelligence in a myriad of different tasks. They are also likely going to fail and be incompliant with human preferences in increasingly subtle ways. Towards the goal of achieving autonomous robots, the robot learning community has made rapid strides in applying machine learning techniques to train robots through data and interaction. This makes the study of how best to audit these algorithms for checking their compatibility with humans, pertinent and urgent. In this paper, we draw inspiration from the AI Safety and Alignment communities and make the case that we need to urgently consider ways in which we can best audit our robot learning algorithms to check for failure modes, and ensure that when operating autonomously, they are indeed behaving in ways that the human algorithm designers intend them to. We believe that this is a challenging problem that will require efforts from the entire robot learning community, and do not attempt to provide a concrete framework for auditing. Instead, we outline high-level guidance and a possible approach towards formulating this framework which we hope will serve as a useful starting point for thinking about auditing in the context of robot learning.

(0 min) AI technologies for Natural Languages have made tremendous progress recently. However, commensurate progress has not been made on Sign Languages, in particular, in recognizing signs as individual words or as complete sentences. We introduce OpenHands, a library where we take four key ideas from the NLP community for low-resource languages and apply them to sign languages for word-level recognition. First, we propose using pose extracted through pretrained models as the standard modality of data to reduce training time and enable efficient inference, and we release standardized pose datasets for 6 different sign languages - American, Argentinian, Chinese, Greek, Indian, and Turkish. Second, we train and release checkpoints of 4 pose-based isolated sign language recognition models across all 6 languages, providing baselines and ready checkpoints for deployment. Third, to address the lack of labelled data, we propose self-supervised pretraining on unlabelled data. We curate and release the largest pose-based pretraining dataset on Indian Sign Language (Indian-SL). Fourth, we compare different pretraining strategies and for the first time establish that pretraining is effective for sign language recognition by demonstrating (a) improved fine-tuning performance especially in low-resource settings, and (b) high crosslingual transfer from Indian-SL to few other sign languages. We open-source all models and datasets in OpenHands with a hope that it makes research in sign languages more accessible, available here at https://github.com/AI4Bharat/OpenHands .

(0 min) The fundamental problem in treatment effect estimation from observational data is confounder identification and balancing. Most of the previous methods realized confounder balancing by treating all observed pre-treatment variables as confounders, ignoring further identifying confounders and non-confounders. In general, not all the observed pre-treatment variables are confounders that refer to the common causes of the treatment and the outcome, some variables only contribute to the treatment and some only contribute to the outcome. Balancing those non-confounders, including instrumental variables and adjustment variables, would generate additional bias for treatment effect estimation. By modeling the different causal relations among observed pre-treatment variables, treatment and outcome, we propose a synergistic learning framework to 1) identify confounders by learning decomposed representations of both confounders and non-confounders, 2) balance confounder with sample re-weighting technique, and simultaneously 3) estimate the treatment effect in observational studies via counterfactual inference. Empirical results on synthetic and real-world datasets demonstrate that the proposed method can precisely decompose confounders and achieve a more precise estimation of treatment effect than baselines.

(0 min) An effective weighting scheme for training samples is essential for learning tasks. Numerous weighting schemes have been proposed. Some schemes take the easy-first mode on samples, whereas some others take the hard-first mode. Naturally, an interesting yet realistic question is raised. Which samples should be learned first given a new learning task, easy or hard? To answer this question, three aspects of research are carried out. First, a high-level unified weighted loss is proposed, providing a more comprehensive view for existing schemes. Theoretical analysis is subsequently conducted and preliminary conclusions are obtained. Second, a flexible weighting scheme is proposed to overcome the defects of existing schemes. The three modes, namely, easy/medium/hard-first, can be flexibly switched in the proposed scheme. Third, a wide range of experiments are conducted to further compare the weighting schemes in different modes. On the basis of these works, reasonable answers are obtained. Factors including prior knowledge and data characteristics determine which samples should be learned first in a learning task.

(0 min) Ranking system is the core part of modern retrieval and recommender systems, where the goal is to rank candidate items given user contexts. Optimizing ranking systems online means that the deployed system can serve user requests, e.g., queries in the web search, and optimize the ranking policy by learning from user interactions, e.g., clicks. Bandit is a general online learning framework and can be used in our optimization task. However, due to the unique features of ranking, there are several challenges in designing bandit algorithms for ranking system optimization. In this dissertation, we study and propose solutions for four challenges in optimizing ranking systems online: effectiveness, safety, nonstationarity, and diversification. First, the effectiveness is related to how fast the algorithm learns from interactions. We study the effective online ranker evaluation task and propose the MergeDTS algorithm to solve the problem effectively. Second, the deployed algorithm should be safe, which means the algorithm only displays reasonable content to user requests. To solve the safe online learning to rank problem, we propose the BubbleRank algorithm. Third, as users change their preferences constantly, the algorithm should handle the nonstationarity. We formulate this nonstationary online learning to rank problem as cascade non-stationary bandits and propose CascadeDUCB and CascadeSWUCB algorithms to solve the problem. Finally, the contents in ranked lists should be diverse. We consider the results diversification task and propose the CascadeHybird algorithm that considers both the item relevance and results diversification when learning from user interactions.

(0 min) Data collected by multiple methods or from multiple sources is called multi-view data. To make full use of the multi-view data, multi-view learning plays an increasingly important role. Traditional multi-view learning methods rely on a large number of labeled and completed multi-view data. However, it is expensive and time-consuming to obtain a large number of labeled multi-view data in real-world applications. Moreover, multi-view data is often incomplete because of data collection failures, self-deficiency, or other reasons. Therefore, we may have to face the problem of fewer labeled and incomplete multi-view data in real application scenarios. In this paper, a transductive semi-supervised incomplete multi-view TSK fuzzy system modeling method (SSIMV_TSK) is proposed to address these challenges. First, in order to alleviate the dependency on labeled data and keep the model interpretable, the proposed method integrates missing view imputation, pseudo label learning of unlabeled data, and fuzzy system modeling into a single process to yield a model with interpretable fuzzy rules. Then, two new mechanisms, i.e. the bidirectional structural preservation of instance and label, as well as the adaptive multiple alignment collaborative learning, are proposed to improve the robustness of the model. The proposed method has the following distinctive characteristics: 1) it can deal with the incomplete and few labeled multi-view data simultaneously; 2) it integrates the missing view imputation and model learning as a single process, which is more efficient than the traditional two-step strategy; 3) attributed to the interpretable fuzzy inference rules, this method is more interpretable. Experimental results on real datasets show that the proposed method significantly outperforms the state-of-the-art methods.

(0 min) We analyze the complexity of learning directed acyclic graphical models from observational data in general settings without specific distributional assumptions. Our approach is information-theoretic and uses a local Markov boundary search procedure in order to recursively construct ancestral sets in the underlying graphical model. Perhaps surprisingly, we show that for certain graph ensembles, a simple forward greedy search algorithm (i.e. without a backward pruning phase) suffices to learn the Markov boundary of each node. This substantially improves the sample complexity, which we show is at most polynomial in the number of nodes. This is then applied to learn the entire graph under a novel identifiability condition that generalizes existing conditions from the literature. As a matter of independent interest, we establish finite-sample guarantees for the problem of recovering Markov boundaries from data. Moreover, we apply our results to the special case of polytrees, for which the assumptions simplify, and provide explicit conditions under which polytrees are identifiable and learnable in polynomial time. We further illustrate the performance of the algorithm, which is easy to implement, in a simulation study. Our approach is general, works for discrete or continuous distributions without distributional assumptions, and as such sheds light on the minimal assumptions required to efficiently learn the structure of directed graphical models from data.

(0 min) Past research and practice have demonstrated that dynamic rerouting framework is effective in mitigating urban traffic congestion and thereby improve urban travel efficiency. It has been suggested that dynamic rerouting could be facilitated using emerging technologies such as fog-computing which offer advantages of low-latency capabilities and information exchange between vehicles and roadway infrastructure. To address this question, this study proposes a two-stage model that combines GAQ (Graph Attention Network - Deep Q Learning) and EBkSP (Entropy Based k Shortest Path) using a fog-cloud architecture, to reroute vehicles in a dynamic urban environment and therefore to improve travel efficiency in terms of travel speed. First, GAQ analyzes the traffic conditions on each road and for each fog area, and then assigns a road index based on the information attention from both local and neighboring areas. Second, EBkSP assigns the route for each vehicle based on the vehicle priority and route popularity. A case study experiment is carried out to investigate the efficacy of the proposed model. At the model training stage, different methods are used to establish the vehicle priorities, and their impact on the results is assessed. Also, the proposed model is tested under various scenarios with different ratios of rerouting and background (non-rerouting) vehicles. The results demonstrate that vehicle rerouting using the proposed model can help attain higher speed and reduces possibility of severe congestion. This result suggests that the proposed model can be deployed by urban transportation agencies for dynamic rerouting and ultimately, to reduce urban traffic congestion.

(0 min) An established way to improve the transferability of black-box evasion attacks is to craft the adversarial examples on a surrogate ensemble model to increase diversity. We argue that transferability is fundamentally related to epistemic uncertainty. Based on a state-of-the-art Bayesian Deep Learning technique, we propose a new method to efficiently build a surrogate by sampling approximately from the posterior distribution of neural network weights, which represents the belief about the value of each parameter. Our extensive experiments on ImageNet and CIFAR-10 show that our approach improves the transfer rates of four state-of-the-art attacks significantly (up to 62.1 percentage points), in both intra-architecture and inter-architecture cases. On ImageNet, our approach can reach 94% of transfer rate while reducing training computations from 11.6 to 2.4 exaflops, compared to an ensemble of independently trained DNNs. Our vanilla surrogate achieves 87.5% of the time higher transferability than 3 test-time techniques designed for this purpose. Our work demonstrates that the way to train a surrogate has been overlooked although it is an important element of transfer-based attacks. We are, therefore, the first to review the effectiveness of several training methods in increasing transferability. We provide new directions to better understand the transferability phenomenon and offer a simple but strong baseline for future work.

(0 min) Robust principal component analysis (RPCA) is a critical tool in modern machine learning, which detects outliers in the task of low-rank matrix reconstruction. In this paper, we propose a scalable and learnable non-convex approach for high-dimensional RPCA problems, which we call Learned Robust PCA (LRPCA). LRPCA is highly efficient, and its free parameters can be effectively learned to optimize via deep unfolding. Moreover, we extend deep unfolding from finite iterations to infinite iterations via a novel feedforward-recurrent-mixed neural network model. We establish the recovery guarantee of LRPCA under mild assumptions for RPCA. Numerical experiments show that LRPCA outperforms the state-of-the-art RPCA algorithms, such as ScaledGD and AltProj, on both synthetic datasets and real-world applications.

(0 min) A popular approach to decompose the neural bases of language consists in correlating, across individuals, the brain responses to different stimuli (e.g. regular speech versus scrambled words, sentences, or paragraphs). Although successful, this `model-free' approach necessitates the acquisition of a large and costly set of neuroimaging data. Here, we show that a model-based approach can reach equivalent results within subjects exposed to natural stimuli. We capitalize on the recently-discovered similarities between deep language models and the human brain to compute the mapping between i) the brain responses to regular speech and ii) the activations of deep language models elicited by modified stimuli (e.g. scrambled words, sentences, or paragraphs). Our model-based approach successfully replicates the seminal study of Lerner et al. (2011), which revealed the hierarchy of language areas by comparing the functional-magnetic resonance imaging (fMRI) of seven subjects listening to 7min of both regular and scrambled narratives. We further extend and precise these results to the brain signals of 305 individuals listening to 4.1 hours of narrated stories. Overall, this study paves the way for efficient and flexible analyses of the brain bases of language.

(0 min) Deep neural networks are known to be vulnerable to adversarially perturbed inputs. A commonly used defense is adversarial training, whose performance is influenced by model capacity. While previous works have studied the impact of varying model width and depth on robustness, the impact of increasing capacity by using learnable parametric activation functions (PAFs) has not been studied. We study how using learnable PAFs can improve robustness in conjunction with adversarial training. We first ask the question: how should we incorporate parameters into activation functions to improve robustness? To address this, we analyze the direct impact of activation shape on robustness through PAFs and observe that activation shapes with positive outputs on negative inputs and with high finite curvature can increase robustness. We combine these properties to create a new PAF, which we call Parametric Shifted Sigmoidal Linear Unit (PSSiLU). We then combine PAFs (including PReLU, PSoftplus and PSSiLU) with adversarial training and analyze robust performance. We find that PAFs optimize towards activation shape properties found to directly affect robustness. Additionally, we find that while introducing only 1-2 learnable parameters into the network, smooth PAFs can significantly increase robustness over ReLU. For instance, when trained on CIFAR-10 with additional synthetic data, PSSiLU improves robust accuracy by 4.54% over ReLU on ResNet-18 and 2.69% over ReLU on WRN-28-10 in the $\ell_{\infty}$ threat model while adding only 2 additional parameters into the network architecture. The PSSiLU WRN-28-10 model achieves 61.96% AutoAttack accuracy, improving over the state-of-the-art robust accuracy on RobustBench (Croce et al., 2020).

(0 min) Graph Representation Learning methods have enabled a wide range of learning problems to be addressed for data that can be represented in graph form. Nevertheless, several real world problems in economy, biology, medicine and other fields raised relevant scaling problems with existing methods and their software implementation, due to the size of real world graphs characterized by millions of nodes and billions of edges. We present GraPE, a software resource for graph processing and random walk based embedding, that can scale with large and high-degree graphs and significantly speed up-computation. GraPE comprises specialized data structures, algorithms, and a fast parallel implementation that displays everal orders of magnitude improvement in empirical space and time complexity compared to state of the art software resources, with a corresponding boost in the performance of machine learning methods for edge and node label prediction and for the unsupervised analysis of graphs.GraPE is designed to run on laptop and desktop computers, as well as on high performance computing clusters

(0 min) The worldwide adoption of cloud data centers (CDCs) has given rise to the ubiquitous demand for hosting application services on the cloud. Further, contemporary data-intensive industries have seen a sharp upsurge in the resource requirements of modern applications. This has led to the provisioning of an increased number of cloud servers, giving rise to higher energy consumption and, consequently, sustainability concerns. Traditional heuristics and reinforcement learning based algorithms for energy-efficient cloud resource management address the scalability and adaptability related challenges to a limited extent. Existing work often fails to capture dependencies across thermal characteristics of hosts, resource consumption of tasks and the corresponding scheduling decisions. This leads to poor scalability and an increase in the compute resource requirements, particularly in environments with non-stationary resource demands. To address these limitations, we propose an artificial intelligence (AI) based holistic resource management technique for sustainable cloud computing called HUNTER. The proposed model formulates the goal of optimizing energy efficiency in data centers as a multi-objective scheduling problem, considering three important models: energy, thermal and cooling. HUNTER utilizes a Gated Graph Convolution Network as a surrogate model for approximating the Quality of Service (QoS) for a system state and generating optimal scheduling decisions. Experiments on simulated and physical cloud environments using the CloudSim toolkit and the COSCO framework show that HUNTER outperforms state-of-the-art baselines in terms of energy consumption, SLA violation, scheduling time, cost and temperature by up to 12, 35, 43, 54 and 3 percent respectively.

(0 min) Normalizing flows have shown great success as general-purpose density estimators. However, many real world applications require the use of domain-specific knowledge, which normalizing flows cannot readily incorporate. We propose embedded-model flows(EMF), which alternate general-purpose transformations with structured layers that embed domain-specific inductive biases. These layers are automatically constructed by converting user-specified differentiable probabilistic models into equivalent bijective transformations. We also introduce gated structured layers, which allow bypassing the parts of the models that fail to capture the statistics of the data. We demonstrate that EMFs can be used to induce desirable properties such as multimodality, hierarchical coupling and continuity. Furthermore, we show that EMFs enable a high performance form of variational inference where the structure of the prior model is embedded in the variational architecture. In our experiments, we show that this approach outperforms state-of-the-art methods in common structured inference problems.

(0 min) Phonocardiogram (PCG) signal analysis is a critical, widely-studied technology to noninvasively analyze the heart's mechanical activity. Through evaluating heart sounds, this technology has been chiefly leveraged as a preliminary solution to automatically diagnose Cardiovascular diseases among adults; however, prenatal tasks such as fetal gender identification have been relatively less studied using fetal Phonocardiography (FPCG). In this work, we apply common PCG signal processing techniques on the gender-tagged Shiraz University Fetal Heart Sounds Database and study the applicability of previously proposed features in classifying fetal gender using both Machine Learning and Deep Learning models. Even though PCG data acquisition's cost-effectiveness and feasibility make it a convenient method of Fetal Heart Rate (FHR) monitoring, the contaminated nature of PCG signals with the noise of various types makes it a challenging modality. To address this problem, we experimented with both static and adaptive noise reduction techniques such as Low-pass filtering, Denoising Autoencoders, and Source Separators. We apply a wide range of previously proposed classifiers to our dataset and propose a novel ensemble method of Fetal Gender Identification (FGI). Our method substantially outperformed the baseline and reached up to 91% accuracy in classifying fetal gender of unseen subjects.

(0 min) Music sentiment transfer is a completely novel task. Sentiment transfer is a natural evolution of the heavily-studied style transfer task, as sentiment transfer is rooted in applying the sentiment of a source to be the new sentiment for a target piece of media; yet compared to style transfer, sentiment transfer has been only scantily studied on images. Music sentiment transfer attempts to apply the high level objective of sentiment transfer to the domain of music. We propose CycleGAN to bridge disparate domains. In order to use the network, we choose to use symbolic, MIDI, data as the music format. Through the use of a cycle consistency loss, we are able to create one-to-one mappings that preserve the content and realism of the source data. Results and literature suggest that the task of music sentiment transfer is more difficult than image sentiment transfer because of the temporal characteristics of music and lack of existing datasets.

(0 min) This paper unifies the multi-focus images fusion (MFIF) and blind super resolution (SR) problems as the multi-focus image super resolution fusion (MFISRF) task, and proposes a novel unified dataset-free unsupervised framework named deep fusion prior (DFP) to address such MFISRF task. DFP consists of SKIPnet network, DoubleReblur focus measurement tactic, decision embedding module and loss functions. In particular, DFP can obtain MFISRF only from two low-resolution inputs without any extent dataset; SKIPnet implementing unsupervised learning via deep image prior is an end-to-end generated network acting as the engine of DFP; DoubleReblur is used to determine the primary decision map without learning but based on estimated PSF and Gaussian kernels convolution; decision embedding module optimizes the decision map via learning; and DFP losses composed of content loss, joint gradient loss and gradient limit loss can obtain high-quality MFISRF results robustly. Experiments have proved that our proposed DFP approaches and even outperforms those state-of-art MFIF and SR method combinations. Additionally, DFP is a general framework, thus its networks and focus measurement tactics can be continuously updated to further improve the MFISRF performance. DFP codes are open source and will be available soon at this http URL

(0 min) When training and evaluating machine learning models on a large number of tasks, it is important to not only look at average task accuracy -- which may be biased by easy or redundant tasks -- but also worst-case accuracy (i.e. the performance on the task with the lowest accuracy). In this work, we show how to use techniques from the distributionally robust optimization (DRO) literature to improve worst-case performance in multitask learning. We highlight several failure cases of DRO when applied off-the-shelf and present an improved method, Lookahead-DRO (L-DRO), which mitigates these issues. The core idea of L-DRO is to anticipate the interaction between tasks during training in order to choose a dynamic re-weighting of the various task losses, which will (i) lead to minimal worst-case loss and (ii) train on as many tasks as possible. After demonstrating the efficacy of L-DRO on a small controlled synthetic setting, we evaluate it on two realistic benchmarks: a multitask version of the CIFAR-100 image classification dataset and a large-scale multilingual language modeling experiment. Our empirical results show that L-DRO achieves a better trade-off between average and worst-case accuracy with little computational overhead compared to several strong baselines.

(0 min) Most existing black-box optimization methods assume that all variables in the system being optimized have equal cost and can change freely at each iteration. However, in many real world systems, inputs are passed through a sequence of different operations or modules, making variables in earlier stages of processing more costly to update. Such structure imposes a cost on switching variables in early parts of a data processing pipeline. In this work, we propose a new algorithm for switch cost-aware optimization called Lazy Modular Bayesian Optimization (LaMBO). This method efficiently identifies the global optimum while minimizing cost through a passive change of variables in early modules. The method is theoretical grounded and achieves vanishing regret when augmented with switching cost. We apply LaMBO to multiple synthetic functions and a three-stage image segmentation pipeline used in a neuroscience application, where we obtain promising improvements over prevailing cost-aware Bayesian optimization algorithms. Our results demonstrate that LaMBO is an effective strategy for black-box optimization that is capable of minimizing switching costs in modular systems.

(0 min) We design an active learning algorithm for cost-sensitive multiclass classification: problems where different errors have different costs. Our algorithm, COAL, makes predictions by regressing to each label's cost and predicting the smallest. On a new example, it uses a set of regressors that perform well on past data to estimate possible costs for each label. It queries only the labels that could be the best, ignoring the sure losers. We prove COAL can be efficiently implemented for any regression family that admits squared loss optimization; it also enjoys strong guarantees with respect to predictive performance and labeling effort. We empirically compare COAL to passive learning and several active learning baselines, showing significant improvements in labeling effort and test cost on real-world datasets.

(0 min) We present principled approaches to train and deploy dyadic neural embedding models at the billion scale, focusing our investigation on the application of semantic product search. When training a dyadic model, one seeks to embed two different types of entities (e.g., queries and documents or users and movies) in a common vector space such that pairs with high relevance are positioned nearby. During inference, given an embedding of one type (e.g., a query or a user), one seeks to retrieve the entities of the other type (e.g., documents or movies, respectively) that are highly relevant. In this work, we show that exploiting the natural structure of real-world datasets helps address both challenges efficiently. Specifically, we model dyadic data as a bipartite graph with edges between pairs with positive associations. We then propose to partition this network into semantically coherent clusters and thus reduce our search space by focusing on a small subset of these partitions for a given input. During training, this technique enables us to efficiently mine hard negative examples while, at inference, we can quickly find the nearest neighbors for a given embedding. We provide offline experimental results that demonstrate the efficacy of our techniques for both training and inference on a billion-scale Amazon.com product search dataset.

(0 min) The need to avoid confident predictions on unfamiliar data has sparked interest in out-of-distribution (OOD) detection. It is widely assumed that Bayesian neural networks (BNNs) are well suited for this task, as the endowed epistemic uncertainty should lead to disagreement in predictions on outliers. In this paper, we question this assumption and show that proper Bayesian inference with function space priors induced by neural networks does not necessarily lead to good OOD detection. To circumvent the use of approximate inference, we start by studying the infinite-width case, where Bayesian inference can be exact due to the correspondence with Gaussian processes. Strikingly, the kernels induced under common architectural choices lead to uncertainties that do not reflect the underlying data generating process and are therefore unsuited for OOD detection. Importantly, we find this OOD behavior to be consistent with the corresponding finite-width networks. Desirable function space properties can be encoded in the prior in weight space, however, this currently only applies to a specified subset of the domain and thus does not inherently extend to OOD data. Finally, we argue that a trade-off between generalization and OOD capabilities might render the application of BNNs for OOD detection undesirable in practice. Overall, our study discloses fundamental problems when naively using BNNs for OOD detection and opens interesting avenues for future research.

(0 min) We propose a novel approach to data-driven modeling of a transient production of oil wells. We apply the transformer-based neural networks trained on the multivariate time series composed of various parameters of oil wells measured during their exploitation. By tuning the machine learning models for a single well (ignoring the effect of neighboring wells) on the open-source field datasets, we demonstrate that transformer outperforms recurrent neural networks with LSTM/GRU cells in the forecasting of the bottomhole pressure dynamics. We apply the transfer learning procedure to the transformer-based surrogate model, which includes the initial training on the dataset from a certain well and additional tuning of the model's weights on the dataset from a target well. Transfer learning approach helps to improve the prediction capability of the model. Next, we generalize the single-well model based on the transformer architecture for multiple wells to simulate complex transient oilfield-level patterns. In other words, we create the global model which deals with the dataset, comprised of the production history from multiple wells, and allows for capturing the well interference resulting in more accurate prediction of the bottomhole pressure or flow rate evolutions for each well under consideration. The developed instruments for a single-well and oilfield-scale modelling can be used to optimize the production process by selecting the operating regime and submersible equipment to increase the hydrocarbon recovery. In addition, the models can be helpful to perform well-testing avoiding costly shut-in operations.

(0 min) We present a fast block direct solver for the unified dynamic simulations of power systems. This solver uses a novel Q-learning based method for task scheduling. Unified dynamic simulations of power systems represent a method in which the electric-mechanical transient, medium-term and long-term dynamic phenomena are organically united. Due to the high rank and large numbers in solving, fast solution of these equations is the key to speeding up the simulation. The sparse systems of simulation contain complex nested block structure, which could be used by the solver to speed up. For the scheduling of blocks and frontals in the solver, we use a learning based task-tree scheduling technique in the framework of Markov Decision Process. That is, we could learn optimal scheduling strategies by offline training on many sample matrices. Then for any systems, the solver would get optimal task partition and scheduling on the learned model. Our learning-based algorithm could help improve the performance of sparse solver, which has been verified in some numerical experiments. The simulation on some large power systems shows that our solver is 2-6 times faster than KLU, which is the state-of-the-art sparse solver for circuit simulation problems.

(0 min) Hierarchical taxonomies are common in many contexts, and they are a very natural structure humans use to organise information. In machine learning, the family of methods that use the 'extra' information is called hierarchical classification. However, applied to audio classification, this remains relatively unexplored. Here we focus on how to integrate the hierarchical information of a problem to learn embeddings representative of the hierarchical relationships. Previously, triplet loss has been proposed to address this problem, however it presents some issues like requiring the careful construction of the triplets, and being limited in the extent of hierarchical information it uses at each iteration. In this work we propose a rank based loss function that uses hierarchical information and translates this into a rank ordering of target distances between the examples. We show that rank based loss is suitable to learn hierarchical representations of the data. By testing on unseen fine level classes we show that this method is also capable of learning hierarchically correct representations of the new classes. Rank based loss has two promising aspects, it is generalisable to hierarchies with any number of levels, and is capable of dealing with data with incomplete hierarchical labels.

(0 min) As a promising tool to navigate in the vast chemical space, artificial intelligence (AI) is leveraged for drug design. From the year 2017 to 2021, the number of applications of several recent AI models (i.e. graph neural network (GNN), recurrent neural network (RNN), variation autoencoder (VAE), generative adversarial network (GAN), flow and reinforcement learning (RL)) in drug design increases significantly. Many relevant literature reviews exist. However, none of them provides an in-depth summary of many applications of the recent AI models in drug design. To complement the existing literature, this survey includes the theoretical development of the previously mentioned AI models and detailed summaries of 42 recent applications of AI in drug design. Concretely, 13 of them leverage GNN for molecular property prediction and 29 of them use RL and/or deep generative models for molecule generation and optimization. In most cases, the focus of the summary is the models, their variants, and modifications for specific tasks in drug design. Moreover, 60 additional applications of AI in molecule generation and optimization are briefly summarized in a table. Finally, this survey provides a holistic discussion of the abundant applications so that the tasks, potential solutions, and challenges in AI-based drug design become evident.

(0 min) Generative diffusion processes are an emerging and effective tool for image and speech generation. In the existing methods, the underlying noise distribution of the diffusion process is Gaussian noise. However, fitting distributions with more degrees of freedom could improve the performance of such generative models. In this work, we investigate other types of noise distribution for the diffusion process. Specifically, we introduce the Denoising Diffusion Gamma Model (DDGM) and show that noise from Gamma distribution provides improved results for image and speech generation. Our approach preserves the ability to efficiently sample state in the training diffusion process while using Gamma noise.

(0 min) A fundamental concept in control theory is that of controllability, where any system state can be reached through an appropriate choice of control inputs. Indeed, a large body of classical and modern approaches are designed for controllable linear dynamical systems. However, in practice, we often encounter systems in which a large set of state variables evolve exogenously and independently of the control inputs; such systems are only \emph{partially controllable}. The focus of this work is on a large class of partially controllable linear dynamical systems, specified by an underlying sparsity pattern. Our main results establish structural conditions and finite-sample guarantees for learning to control such systems. In particular, our structural results characterize those state variables which are irrelevant for optimal control, an analysis which departs from classical control techniques. Our algorithmic results adapt techniques from high-dimensional statistics -- specifically soft-thresholding and semiparametric least-squares -- to exploit the underlying sparsity pattern in order to obtain finite-sample guarantees that significantly improve over those based on certainty-equivalence. We also corroborate these theoretical improvements over certainty-equivalent control through a simulation study.

(0 min) Training Deep Neural Networks (DNNs) is a widely popular workload in both enterprises and cloud data centers. Existing schedulers for DNN training consider GPU as the dominant resource, and allocate other resources such as CPU and memory proportional to the number of GPUs requested by the job. Unfortunately, these schedulers do not consider the impact of a job's sensitivity to allocation of CPU, memory, and storage resources. In this work, we propose Synergy, a resource-sensitive scheduler for shared GPU clusters. Synergy infers the sensitivity of DNNs to different resources using optimistic profiling; some jobs might benefit from more than the GPU-proportional allocation and some jobs might not be affected by less than GPU-proportional allocation. Synergy performs such multi-resource workload-aware assignments across a set of jobs scheduled on shared multi-tenant clusters using a new near-optimal online algorithm. Our experiments show that workload-aware CPU and memory allocations can improve average JCT up to 3.4x when compared to traditional GPU-proportional scheduling.

(0 min) The goal of Domain Generation Algorithm (DGA) detection is to recognize infections with bot malware and is often done with help of Machine Learning approaches that classify non-resolving Domain Name System (DNS) traffic and are trained on possibly sensitive data. In parallel, the rise of privacy research in the Machine Learning world leads to privacy-preserving measures that are tightly coupled with a deep learning model's architecture or training routine, while non deep learning approaches are commonly better suited for the application of privacy-enhancing methods outside the actual classification module. In this work, we aim to measure the privacy capability of the feature extractor of feature-based DGA detector FANCI (Feature-based Automated Nxdomain Classification and Intelligence). Our goal is to assess whether a data-rich adversary can learn an inverse mapping of FANCI's feature extractor and thereby reconstruct domain names from feature vectors. Attack success would pose a privacy threat to sharing FANCI's feature representation, while the opposite would enable this representation to be shared without privacy concerns. Using three real-world data sets, we train a recurrent Machine Learning model on the reconstruction task. Our approaches result in poor reconstruction performance and we attempt to back our findings with a mathematical review of the feature extraction process. We thus reckon that sharing FANCI's feature representation does not constitute a considerable privacy leakage.

(0 min) Many popular adaptive gradient methods such as Adam and RMSProp rely on an exponential moving average (EMA) to normalize their stepsizes. While the EMA makes these methods highly responsive to new gradient information, recent research has shown that it also causes divergence on at least one convex optimization problem. We propose a novel method called Expectigrad, which adjusts stepsizes according to a per-component unweighted mean of all historical gradients and computes a bias-corrected momentum term jointly between the numerator and denominator. We prove that Expectigrad cannot diverge on every instance of the optimization problem known to cause Adam to diverge. We also establish a regret bound in the general stochastic nonconvex setting that suggests Expectigrad is less susceptible to gradient variance than existing methods are. Testing Expectigrad on several high-dimensional machine learning tasks, we find it often performs favorably to state-of-the-art methods with little hyperparameter tuning.

(0 min) When deployed in the real world, machine learning models inevitably encounter changes in the data distribution, and certain -- but not all -- distribution shifts could result in significant performance degradation. In practice, it may make sense to ignore benign shifts, under which the performance of a deployed model does not degrade substantially, making interventions by a human expert (or model retraining) unnecessary. While several works have developed tests for distribution shifts, these typically either use non-sequential methods, or detect arbitrary shifts (benign or harmful), or both. We argue that a sensible method for firing off a warning has to both (a) detect harmful shifts while ignoring benign ones, and (b) allow continuous monitoring of model performance without increasing the false alarm rate. In this work, we design simple sequential tools for testing if the difference between source (training) and target (test) distributions leads to a significant drop in a risk function of interest, like accuracy or calibration. Recent advances in constructing time-uniform confidence sequences allow efficient aggregation of statistical evidence accumulated during the tracking process. The designed framework is applicable in settings where (some) true labels are revealed after the prediction is performed, or when batches of labels become available in a delayed fashion. We demonstrate the efficacy of the proposed framework through an extensive empirical study on a collection of simulated and real datasets.

(0 min) The processing of signals supported on non-Euclidean domains has attracted large interest in the last years. Thus far, such non-Euclidean domains have been abstracted primarily as graphs with signals supported on the nodes, though recently the processing of signals on more general structures such as simplicial complexes has also been considered. In this paper, we give an introduction to signal processing on (abstract) regular cell complexes, which provide a unifying framework encompassing graphs, simplicial complexes, cubical complexes and various meshes as special cases. We discuss how appropriate Hodge Laplacians for these cell complexes can be derived. These Hodge Laplacians enable the construction of convolutional filters, which can be employed in linear filtering and non-linear filtering via neural networks defined on cell complexes.

(0 min) Controllable music generation with deep generative models has become increasingly reliant on disentanglement learning techniques. However, current disentanglement metrics, such as mutual information gap (MIG), are often inadequate and misleading when used for evaluating latent representations in the presence of interdependent semantic attributes often encountered in real-world music datasets. In this work, we propose a dependency-aware information metric as a drop-in replacement for MIG that accounts for the inherent relationship between semantic attributes.

(0 min) "The power of a generalization system follows directly from its biases" (Mitchell 1980). Today, CNNs are incredibly powerful generalisation systems -- but to what degree have we understood how their inductive bias influences model decisions? We here attempt to disentangle the various aspects that determine how a model decides. In particular, we ask: what makes one model decide differently from another? In a meticulously controlled setting, we find that (1.) irrespective of the network architecture or objective (e.g. self-supervised, semi-supervised, vision transformers, recurrent models) all models end up with a similar decision boundary. (2.) To understand these findings, we analysed model decisions on the ImageNet validation set from epoch to epoch and image by image. We find that the ImageNet validation set, among others, suffers from dichotomous data difficulty (DDD): For the range of investigated models and their accuracies, it is dominated by 46.0% "trivial" and 11.5% "impossible" images (beyond label errors). Only 42.5% of the images could possibly be responsible for the differences between two models' decision boundaries. (3.) Only removing the "impossible" and "trivial" images allows us to see pronounced differences between models. (4.) Humans are highly accurate at predicting which images are "trivial" and "impossible" for CNNs (81.4%). This implies that in future comparisons of brains, machines and behaviour, much may be gained from investigating the decisive role of images and the distribution of their difficulties.

(0 min) Adversarial attacks can generate adversarial inputs by applying small but intentionally worst-case perturbations to samples from the dataset, which leads to even state-of-the-art deep neural networks outputting incorrect answers with high confidence. Hence, some adversarial defense techniques are developed to improve the security and robustness of the models and avoid them being attacked. Gradually, a game-like competition between attackers and defenders formed, in which both players would attempt to play their best strategies against each other while maximizing their own payoffs. To solve the game, each player would choose an optimal strategy against the opponent based on the prediction of the opponent's strategy choice. In this work, we are on the defensive side to apply game-theoretic approaches on defending against attacks. We use two randomization methods, random initialization and stochastic activation pruning, to create diversity of networks. Furthermore, we use one denoising technique, super resolution, to improve models' robustness by preprocessing images before attacks. Our experimental results indicate that those three methods can effectively improve the robustness of deep-learning neural networks.

(0 min) Solutions to the Traveling Salesperson Problem (TSP) have practical applications to processes in transportation, logistics, and automation, yet must be computed with minimal delay to satisfy the real-time nature of the underlying tasks. However, solving large TSP instances quickly without sacrificing solution quality remains challenging for current approximate algorithms. To close this gap, we present a hybrid data-driven approach for solving the TSP based on Graph Neural Networks (GNNs) and Guided Local Search (GLS). Our model predicts the regret of including each edge of the problem graph in the solution; GLS uses these predictions in conjunction with the original problem graph to find solutions. Our experiments demonstrate that this approach converges to optimal solutions at a faster rate than state-of-the-art learning-based approaches and non-learning GLS algorithms for the TSP, notably finding optimal solutions to 96% of the 50-node problem set, 7% more than the next best benchmark, and to 20% of the 100-node problem set, 4.5x more than the next best benchmark. When generalizing from 20-node problems to the 100-node problem set, our approach finds solutions with an average optimality gap of 2.5%, a 10x improvement over the next best learning-based benchmark.

(0 min) DETR is a recently proposed Transformer-based method which views object detection as a set prediction problem and achieves state-of-the-art performance but demands extra-long training time to converge. In this paper, we investigate the causes of the optimization difficulty in the training of DETR. Our examinations reveal several factors contributing to the slow convergence of DETR, primarily the issues with the Hungarian loss and the Transformer cross-attention mechanism. To overcome these issues we propose two solutions, namely, TSP-FCOS (Transformer-based Set Prediction with FCOS) and TSP-RCNN (Transformer-based Set Prediction with RCNN). Experimental results show that the proposed methods not only converge much faster than the original DETR, but also significantly outperform DETR and other baselines in terms of detection accuracy.

(0 min) The Schr\"odinger bridge problem (SBP) finds the most likely stochastic evolution between two probability distributions given a prior stochastic evolution. As well as applications in the natural sciences, problems of this kind have important applications in machine learning such as dataset alignment and hypothesis testing. Whilst the theory behind this problem is relatively mature, scalable numerical recipes to estimate the Schr\"odinger bridge remain an active area of research. We prove an equivalence between the SBP and maximum likelihood estimation enabling direct application of successful machine learning techniques. We propose a numerical procedure to estimate SBPs using Gaussian process and demonstrate the practical usage of our approach in numerical simulations and experiments.

(0 min) Data augmentation has been an indispensable tool to improve the performance of deep neural networks, however the augmentation can hardly transfer among different tasks and datasets. Consequently, a recent trend is to adopt AutoML technique to learn proper augmentation policy without extensive hand-crafted tuning. In this paper, we propose an efficient differentiable search algorithm called Direct Differentiable Augmentation Search (DDAS). It exploits meta-learning with one-step gradient update and continuous relaxation to the expected training loss for efficient search. Our DDAS can achieve efficient augmentation search without relying on approximations such as Gumbel Softmax or second order gradient approximation. To further reduce the adverse effect of improper augmentations, we organize the search space into a two level hierarchy, in which we first decide whether to apply augmentation, and then determine the specific augmentation policy. On standard image classification benchmarks, our DDAS achieves state-of-the-art performance and efficiency tradeoff while reducing the search cost dramatically, e.g. 0.15 GPU hours for CIFAR-10. In addition, we also use DDAS to search augmentation for object detection task and achieve comparable performance with AutoAugment, while being 1000x faster.

(0 min) Perceived signals in real-world scenarios are usually high-dimensional and noisy, and finding and using their representation that contains essential and sufficient information required by downstream decision-making tasks will help improve computational efficiency and generalization ability in the tasks. In this paper, we focus on partially observable environments and propose to learn a minimal set of state representations that capture sufficient information for decision-making, termed \textit{Action-Sufficient state Representations} (ASRs). We build a generative environment model for the structural relationships among variables in the system and present a principled way to characterize ASRs based on structural constraints and the goal of maximizing cumulative reward in policy learning. We then develop a structured sequential Variational Auto-Encoder to estimate the environment model and extract ASRs. Our empirical results on CarRacing and VizDoom demonstrate a clear advantage of learning and using ASRs for policy learning. Moreover, the estimated environment model and ASRs allow learning behaviors from imagined outcomes in the compact latent space to improve sample efficiency.

(0 min) Given a set of data points belonging to the convex hull of a set of vertices, a key problem in data analysis and machine learning is to estimate these vertices in the presence of noise. Many algorithms have been developed under the assumption that there is at least one nearby data point to each vertex; two of the most widely used ones are vertex component analysis (VCA) and the successive projection algorithm (SPA). This assumption is known as the pure-pixel assumption in blind hyperspectral unmixing, and as the separability assumption in nonnegative matrix factorization. More recently, Bhattacharyya and Kannan (ACM-SIAM Symposium on Discrete Algorithms, 2020) proposed an algorithm for learning a latent simplex (ALLS) that relies on the assumption that there is more than one nearby data point for each vertex. In that scenario, ALLS is probalistically more robust to noise than algorithms based on the separability assumption. In this paper, inspired by ALLS, we propose smoothed VCA (SVCA) and smoothed SPA (SSPA) that generalize VCA and SPA by assuming the presence of several nearby data points to each vertex. We illustrate the effectiveness of SVCA and SSPA over VCA, SPA and ALLS on synthetic data sets, and on the unmixing of hyperspectral images.

(0 min) Fundamental knowledge in activity recognition of individuals with motor disorders such as Parkinson's disease (PD) has been primarily limited to detection of steady-state/static tasks (sitting, standing, walking). To date, identification of non-steady-state locomotion on uneven terrains (stairs, ramps) has not received much attention. Furthermore, previous research has mainly relied on data from a large number of body locations which could adversely affect user convenience and system performance. Here, individuals with mild stages of PD and healthy subjects performed non-steady-state circuit trials comprising stairs, ramp, and changes of direction. An offline analysis using a linear discriminant analysis (LDA) classifier and a Long-Short Term Memory (LSTM) neural network was performed for task recognition. The performance of accelerographic and gyroscopic information from varied lower/upper-body segments were tested across a set of user-independent and user-dependent training paradigms. Comparing the F1 score of a given signal across classifiers showed improved performance using LSTM compared to LDA. Using LSTM, even a subset of information (e.g., feet data) in subject-independent training appeared to provide F1 score > 0.8. However, employing LDA was shown to be at the expense of being limited to using a subject-dependent training and/or biomechanical data from multiple body locations. The findings could inform a number of applications in the field of healthcare monitoring and developing advanced lower-limb assistive devices by providing insights into classification schemes capable of handling non-steady-state and unstructured locomotion in individuals with mild Parkinson's disease.

(0 min) Context: Modelling satellite galaxy abundance $N_s$ in Galaxy Clusters (GCs) is a key element in modelling the Halo Occupation Distribution (HOD), which itself is a powerful tool to connect observational studies with numerical simulations. Aims: To study the impact of cosmological parameters on satellite abundance both in cosmological simulations and in mock observations. Methods: We build an emulator (HODEmu, \url{https://github.com/aragagnin/HODEmu/}) of satellite abundance based on cosmological parameters $\Omega_m, \Omega_b, \sigma_8, h_0$ and redshift $z.$ We train our emulator using \magneticum hydrodynamic simulations that span 15 different cosmologies, each over $4$ redshift slices between $0<z<0.5,$ and for each setup we fit normalisation $A$, log-slope $\beta$ and Gaussian fractional-scatter $\sigma$ of the $N_s-M$ relation. The emulator is based on multi-variate output Gaussian Process Regression (GPR). Results: We find that $A$ and $\beta$ depend on cosmological parameters, even if weakly, especially on $\Omega_m,$ $\Omega_b.$ This dependency can explain some discrepancies found in literature between satellite HOD of different cosmological simulations (Magneticum, Illustris, BAHAMAS). We also show that satellite abundance cosmology dependency differs between full-physics (FP) simulations, dark-matter only (DMO), and non-radiative simulations. Conclusions: This work provides a preliminary calibration of the cosmological dependency of the satellite abundance of high mass halos, and we showed that modelling HOD with cosmological parameters is necessary to interpret satellite abundance, and we showed the importance of using FP simulations in modelling this dependency.

(0 min) We consider a stochastic multi-armed bandit setting where feedback is limited by a (possibly time-dependent) budget, and reward must be actively inquired for it to be observed. Previous works on this setting assumed a strict feedback budget and focused on not violating this constraint while providing problem-independent regret guarantees. In this work, we provide problem-dependent guarantees on both the regret and the asked feedback. In particular, we derive problem-dependent lower bounds on the required feedback and show that there is a fundamental difference between problems with a unique and multiple optimal arms. Furthermore, we present a new algorithm called BuFALU for which we derive problem-dependent regret and cumulative feedback bounds. Notably, we show that BuFALU naturally adapts to the number of optimal arms.

(0 min) This paper presents a novel cross-coupling capacitor processing unit (C3PU) that supports analog-mixed signal in memory computing to perform multiply-and-accumulate (MAC) operations. The C3PU consists of a capacitive unit, a CMOS transistor, and a voltage-to-time converter (VTC). The capacitive unit serves as a computational element that holds the multiplier operand and performs multiplication once the multiplicand is applied at the terminal. The multiplicand is the input voltage that is converted to a pulse width signal using a low power VTC. The transistor transfers this multiplication where a voltage level is generated. A demonstrator of 5x4 C3PU array that is capable of implementing 4 MAC units is presented. The design has been verified using Monte Carlo simulation in 65 nm technology. The 5x4 C3PU consumed energy of 66.4 fJ/MAC at 0.3 V voltage supply with an error of 5.7%. The proposed unit achieves lower energy and occupies a smaller area by 3.4x and 3.6x, respectively, with similar error value when compared to a digital-based 8x4-bit fixed point MAC unit. The C3PU has been utilized through an iris fower classification utilizing an artificial neural network which achieved a 90% classification accuracy compared to ideal accuracy of 96.67% using MATLAB.

(0 min) We consider the task of visual indoor exploration with multiple agents, where the agents need to cooperatively explore the entire indoor region using as few steps as possible. Classical planning-based methods often suffer from particularly expensive computation at each inference step and a limited expressiveness of cooperation strategy. By contrast, reinforcement learning (RL) has become a trending paradigm for tackling this challenge due to its modeling capability of arbitrarily complex strategies and minimal inference overhead. We extend the state-of-the-art single-agent RL solution, Active Neural SLAM (ANS), to the multi-agent setting by introducing a novel RL-based global-goal planner, Spatial Coordination Planner (SCP), which leverages spatial information from each individual agent in an end-to-end manner and effectively guides the agents to navigate towards different spatial goals with high exploration efficiency. SCP consists of a transformer-based relation encoder to capture intra-agent interactions and a spatial action decoder to produce accurate goals. In addition, we also implement a few multi-agent enhancements to process local information from each agent for an aligned spatial representation and more precise planning. Our final solution, Multi-Agent Active Neural SLAM (MAANS), combines all these techniques and substantially outperforms 4 different planning-based methods and various RL baselines in the photo-realistic physical testbed, Habitat.

(0 min) We present work on summarising deliberative processes for non-English languages. Unlike commonly studied datasets, such as news articles, this deliberation dataset reflects difficulties of combining multiple narratives, mostly of poor grammatical quality, in a single text. We report an extensive evaluation of a wide range of abstractive summarisation models in combination with an off-the-shelf machine translation model. Texts are translated into English, summarised, and translated back to the original language. We obtain promising results regarding the fluency, consistency and relevance of the summaries produced. Our approach is easy to implement for many languages for production purposes by simply changing the translation model.

(0 min) Gaussian processes are widely used for the analysis of spatial data due to their nonparametric flexibility and ability to quantify uncertainty, and recently developed scalable approximations have facilitated application to massive datasets. For multivariate outcomes, linear models of coregionalization combine dimension reduction with spatial correlation. However, their real-valued latent factors and loadings are difficult to interpret because, unlike nonnegative models, they do not recover a parts-based representation. We present nonnegative spatial factorization (NSF), a spatially-aware probabilistic dimension reduction model that naturally encourages sparsity. We compare NSF to real-valued spatial factorizations such as MEFISTO and nonspatial dimension reduction methods using simulations and high-dimensional spatial transcriptomics data. NSF identifies generalizable spatial patterns of gene expression. Since not all patterns of gene expression are spatial, we also propose a hybrid extension of NSF that combines spatial and nonspatial components, enabling quantification of spatial importance for both observations and features. A TensorFlow implementation of NSF is available from https://github.com/willtownes/nsf-paper .

(0 min) The advancement of technology has progressed faster than any other field in the world and with the development of these new technologies, it is important to make sure that these tools can be used by everyone, including people with disabilities. Accessibility options in computing devices help ensure that everyone has the same access to advanced technologies. Unfortunately, for those who require more unique and sometimes challenging accommodations, such as people with Amyotrophic lateral sclerosis ( ALS), the most commonly used accessibility features are simply not enough. While assistive technology for those with ALS does exist, it requires multiple peripheral devices that can become quite expensive collectively. The purpose of this paper is to suggest a more affordable and readily available option for ALS assistive technology that can be implemented on a smartphone or tablet.

(0 min) In this paper, we introduce the Label-Aware Ranked loss, a novel metric loss function. Compared to the state-of-the-art Deep Metric Learning losses, this function takes advantage of the ranked ordering of the labels in regression problems. To this end, we first show that the loss minimises when datapoints of different labels are ranked and laid at uniform angles between each other in the embedding space. Then, to measure its performance, we apply the proposed loss on a regression task of people counting with a short-range radar in a challenging scenario, namely a vehicle cabin. The introduced approach improves the accuracy as well as the neighboring labels accuracy up to 83.0% and 99.9%: An increase of 6.7%and 2.1% on state-of-the-art methods, respectively.

(0 min) Implicit deep learning has received increasing attention recently due to the fact that it generalizes the recursive prediction rules of many commonly used neural network architectures. Its prediction rule is provided implicitly based on the solution of an equilibrium equation. Although a line of recent empirical studies has demonstrated its superior performances, the theoretical understanding of implicit neural networks is limited. In general, the equilibrium equation may not be well-posed during the training. As a result, there is no guarantee that a vanilla (stochastic) gradient descent (SGD) training nonlinear implicit neural networks can converge. This paper fills the gap by analyzing the gradient flow of Rectified Linear Unit (ReLU) activated implicit neural networks. For an $m$-width implicit neural network with ReLU activation and $n$ training samples, we show that a randomly initialized gradient descent converges to a global minimum at a linear rate for the square loss function if the implicit neural network is \textit{over-parameterized}. It is worth noting that, unlike existing works on the convergence of (S)GD on finite-layer over-parameterized neural networks, our convergence results hold for implicit neural networks, where the number of layers is \textit{infinite}.

(0 min) Differential equations play a pivotal role in modern world ranging from science, engineering, ecology, economics and finance where these can be used to model many physical systems and processes. In this paper, we study two mathematical models of a drug assimilation in the human system using Physics Informed Neural Networks (PINNs). In the first model, we consider the case of single dose of drug in the human system and in the second case, we consider the course of this drug taken at regular intervals. We have used the compartment diagram to model these cases. The resulting differential equations are solved using PINN, where we employ a feed forward multilayer perceptron as function approximator and the network parameters are tuned for minimum error. Further, the network is trained by finding the gradient of the error function with respect to the network parameters. We have employed DeepXDE, a python library for PINNs, to solve the simultaneous first order differential equations describing the two models of drug assimilation. The results show high degree of accuracy between the exact solution and the predicted solution as much as the resulting error reaches10^(-11) for the first model and 10^(-8) for the second model. This validates the use of PINN in solving any dynamical system.

(0 min) Personalized medicine, a paradigm of medicine tailored to a patient's characteristics, is an increasingly attractive field in health care. An important goal of personalized medicine is to identify a subgroup of patients, based on baseline covariates, that benefits more from the targeted treatment than other comparative treatments. Most of the current subgroup identification methods only focus on obtaining a subgroup with an enhanced treatment effect without paying attention to subgroup size. Yet, a clinically meaningful subgroup learning approach should identify the maximum number of patients who can benefit from the better treatment. In this paper, we present an optimal subgroup selection rule (SSR) that maximizes the number of selected patients, and in the meantime, achieves the pre-specified clinically meaningful mean outcome, such as the average treatment effect. We derive two equivalent theoretical forms of the optimal SSR based on the contrast function that describes the treatment-covariates interaction in the outcome. We further propose a ConstrAined PolIcy Tree seArch aLgorithm (CAPITAL) to find the optimal SSR within the interpretable decision tree class. The proposed method is flexible to handle multiple constraints that penalize the inclusion of patients with negative treatment effects, and to address time to event data using the restricted mean survival time as the clinically interesting mean outcome. Extensive simulations, comparison studies, and real data applications are conducted to demonstrate the validity and utility of our method.

(0 min) We consider a variant of the novel contextual bandit problem with corrupted context, which we call the contextual bandit problem with corrupted context and action correlation, where actions exhibit a relationship structure that can be exploited to guide the exploration of viable next decisions. Our setting is primarily motivated by adaptive mobile health interventions and related applications, where users might transitions through different stages requiring more targeted action selection approaches. In such settings, keeping user engagement is paramount for the success of interventions and therefore it is vital to provide relevant recommendations in a timely manner. The context provided by users might not always be informative at every decision point and standard contextual approaches to action selection will incur high regret. We propose a meta-algorithm using a referee that dynamically combines the policies of a contextual bandit and multi-armed bandit, similar to previous work, as wells as a simple correlation mechanism that captures action to action transition probabilities allowing for more efficient exploration of time-correlated actions. We evaluate empirically the performance of said algorithm on a simulation where the sequence of best actions is determined by a hidden state that evolves in a Markovian manner. We show that the proposed meta-algorithm improves upon regret in situations where the performance of both policies varies such that one is strictly superior to the other for a given time period. To demonstrate that our setting has relevant practical applicability, we evaluate our method on several real world data sets, clearly showing better empirical performance compared to a set of simple algorithms.

(0 min) Differentially Private (DP) learning has seen limited success for building large deep learning models of text, and attempts at straightforwardly applying Differentially Private Stochastic Gradient Descent (DP-SGD) to NLP tasks have resulted in large performance drops and high computational overhead. We show that this performance drop can be mitigated with (1) the use of large pretrained models; (2) hyperparameters that suit DP optimization; and (3) fine-tuning objectives aligned with the pretraining procedure. With these factors set right, we obtain private NLP models that outperform state-of-the-art private training approaches and strong non-private baselines -- by directly fine-tuning pretrained models with DP optimization on moderately-sized corpora. To address the computational challenge of running DP-SGD with large Transformers, we propose a memory saving technique that allows clipping in DP-SGD to run without instantiating per-example gradients for any layer in the model. The technique enables privately training Transformers with almost the same memory cost as non-private training at a modest run-time overhead. Contrary to conventional wisdom that DP optimization fails at learning high-dimensional models (due to noise that scales with dimension) empirical results reveal that private learning with pretrained models tends to not suffer from dimension-dependent performance degradation.

(0 min) Automatic Speech Recognition (ASR) is gaining special interest in Air Traffic Control (ATC). ASR allows transcribing the communications between air traffic controllers (ATCOs) and pilots. These transcriptions are used to extract ATC command types and named entities such as aircraft callsigns. One common problem is when the Speech Activity Detection (SAD) or diarization system fails and then two or more single speaker segments are in the same recording, jeopardizing the overall system's performance. We developed a system that combines the segmentation of a SAD module with a BERT-based model that performs Speaker Change Detection (SCD) and Speaker Role Identification (SRI) based on ASR transcripts (i.e., diarization + SRI). This research demonstrates on a real-life ATC test set that performing diarization directly on textual data surpass acoustic level diarization. The proposed model reaches up to ~0.90/~0.95 F1-score on ATCO/pilot for SRI on several test sets. The text-based diarization system brings a 27% relative improvement on Diarization Error Rate (DER) compared to standard acoustic-based diarization. These results were on ASR transcripts of a challenging ATC test set with an estimated ~13% word error rate, validating the approach's robustness even on noisy ASR transcripts.

(0 min) Reasoning at multiple levels of temporal abstraction is one of the key attributes of intelligence. In reinforcement learning, this is often modeled through temporally extended courses of actions called options. Options allow agents to make predictions and to operate at different levels of abstraction within an environment. Nevertheless, approaches based on the options framework often start with the assumption that a reasonable set of options is known beforehand. When this is not the case, there are no definitive answers for which options one should consider. In this paper, we argue that the successor representation (SR), which encodes states based on the pattern of state visitation that follows them, can be seen as a natural substrate for the discovery and use of temporal abstractions. To support our claim, we take a big picture view of recent results, showing how the SR can be used to discover options that facilitate either temporally-extended exploration or planning. We cast these results as instantiations of a general framework for option discovery in which the agent's representation is used to identify useful options, which are then used to further improve its representation. This results in a virtuous, never-ending, cycle in which both the representation and the options are constantly refined based on each other. Beyond option discovery itself, we discuss how the SR allows us to augment a set of options into a combinatorially large counterpart without additional learning. This is achieved through the combination of previously learned options. Our empirical evaluation focuses on options discovered for temporally-extended exploration and on the use of the SR to combine them. The results of our experiments shed light on design decisions involved in the definition of options and demonstrate the synergy of different methods based on the SR, such as eigenoptions and the option keyboard.

(0 min) Self-supervised learning holds promise to revolutionize molecule property prediction - a central task to drug discovery and many more industries - by enabling data efficient learning from scarce experimental data. Despite significant progress, non-pretrained methods can be still competitive in certain settings. We reason that architecture might be a key bottleneck. In particular, enriching the backbone architecture with domain-specific inductive biases has been key for the success of self-supervised learning in other domains. In this spirit, we methodologically explore the design space of the self-attention mechanism tailored to molecular data. We identify a novel variant of self-attention adapted to processing molecules, inspired by the relative self-attention layer, which involves fusing embedded graph and distance relationships between atoms. Our main contribution is Relative Molecule Attention Transformer (R-MAT): a novel Transformer-based model based on the developed self-attention layer that achieves state-of-the-art or very competitive results across a~wide range of molecule property prediction tasks.

(0 min) Public transport agencies use social media as an essential tool for communicating mobility incidents to passengers. However, while the short term, day-to-day information about transport phenomena is usually posted in social media with low latency, its availability is short term as the content is rarely made an aggregated form. Social media communication of transport phenomena usually lacks GIS annotations as most social media platforms do not allow attaching non-POI GPS coordinates to posts. As a result, the analysis of transport phenomena information is minimal. We collected three years of social media posts of a polish public transport company with user comments. Through exploration, we infer a six-class transport information typology. We successfully build an information type classifier for social media posts, detect stop names in posts, and relate them to GPS coordinates, obtaining a spatial understanding of long-term aggregated phenomena. We show that our approach enables citizen science and use it to analyze the impact of three years of infrastructure incidents on passenger mobility, and the sentiment and reaction scale towards each of the events. All these results are achieved for Polish, an under-resourced language when it comes to spatial language understanding, especially in social media contexts. To improve the situation, we released two of our annotated data sets: social media posts with incident type labels and matched stop names and social media comments with the annotated sentiment. We also opensource the experimental codebase.

(0 min) The uncertainty of the energy generated by photovoltaic systems incurs an additional cost for a guaranteed, reliable supply of energy (i.e., energy storage). This investigation aims to decrease the additional cost by introducing probabilistic multi-task intra-hour solar forecasting (feasible in real time applications) to increase the penetration of photovoltaic systems in power grids. The direction of moving clouds is estimated in consecutive sequences of sky images by extracting features of cloud dynamics with the objective of forecasting the global solar irradiance that reaches photovoltaic systems. The sky images are acquired using a low-cost infrared sky imager mounted on a solar tracker. The solar forecasting algorithm is based on kernel learning methods, and uses the clear sky index as predictor and features extracted from clouds as feature vectors. The proposed solar forecasting algorithm achieved 16.45\% forecasting skill 8 minutes ahead with a resolution of 15 seconds. In contrast, previous work reached 15.4\% forecasting skill with the resolution of 1 minute. Therefore, this solar forecasting algorithm increases the performances with respect to the state-of-the-art, providing grid operators with the capability of managing the inherent uncertainties of power grids with a high penetration of photovoltaic systems.

(0 min) Phishing emails that appear legitimate lure people into clicking on the attached malicious links or documents. Increasingly more sophisticated phishing campaigns in recent years necessitate a more adaptive detection system other than traditional signature-based methods. In this regard, natural language processing (NLP) with deep neural networks (DNNs) is adopted for knowledge acquisition from a large number of emails. However, such sensitive daily communications containing personal information are difficult to collect on a server for centralized learning in real life due to escalating privacy concerns. To this end, we propose a decentralized phishing email detection method called the Federated Phish Bowl (FPB) leveraging federated learning and long short-term memory (LSTM). FPB allows common knowledge representation and sharing among different clients through the aggregation of trained models to safeguard the email security and privacy. A recent phishing email dataset was collected from an intergovernmental organization to train the model. Moreover, we evaluated the model performance based on various assumptions regarding the total client number and the level of data heterogeneity. The comprehensive experimental results suggest that FPB is robust to a continually increasing client number and various data heterogeneity levels, retaining a detection accuracy of 0.83 and protecting the privacy of sensitive email communications.

(0 min) We study the problem of robotic stacking with objects of complex geometry. We propose a challenging and diverse set of such objects that was carefully designed to require strategies beyond a simple "pick-and-place" solution. Our method is a reinforcement learning (RL) approach combined with vision-based interactive policy distillation and simulation-to-reality transfer. Our learned policies can efficiently handle multiple object combinations in the real world and exhibit a large variety of stacking skills. In a large experimental study, we investigate what choices matter for learning such general vision-based agents in simulation, and what affects optimal transfer to the real robot. We then leverage data collected by such policies and improve upon them with offline RL. A video and a blog post of our work are provided as supplementary material.

(0 min) Surface code error correction offers a highly promising pathway to achieve scalable fault-tolerant quantum computing. When operated as stabilizer codes, surface code computations consist of a syndrome decoding step where measured stabilizer operators are used to determine appropriate corrections for errors in physical qubits. Decoding algorithms have undergone substantial development, with recent work incorporating machine learning (ML) techniques. Despite promising initial results, the ML-based syndrome decoders are still limited to small scale demonstrations with low latency and are incapable of handling surface codes with boundary conditions and various shapes needed for lattice surgery and braiding. Here, we report the development of an artificial neural network (ANN) based scalable and fast syndrome decoder capable of decoding surface codes of arbitrary shape and size with data qubits suffering from the depolarizing error model. Based on rigorous training over 50 million random quantum error instances, our ANN decoder is shown to work with code distances exceeding 1000 (more than 4 million physical qubits), which is the largest ML-based decoder demonstration to-date. The established ANN decoder demonstrates an execution time in principle independent of code distance, implying that its implementation on dedicated hardware could potentially offer surface code decoding times of O($\mu$sec), commensurate with the experimentally realisable qubit coherence times. With the anticipated scale-up of quantum processors within the next decade, their augmentation with a fast and scalable syndrome decoder such as developed in our work is expected to play a decisive role towards experimental implementation of fault-tolerant quantum information processing.

(0 min) We study the problem of recovering the structure underlying large Gaussian graphical models or, more generally, partial correlation graphs. In high-dimensional problems it is often too costly to store the entire sample covariance matrix. We propose a new input model in which one can query single entries of the covariance matrix. We prove that it is possible to recover the support of the inverse covariance matrix with low query and computational complexity. Our algorithms work in a regime when this support is represented by tree-like graphs and, more generally, for graphs of small treewidth. Our results demonstrate that for large classes of graphs, the structure of the corresponding partial correlation graphs can be determined much faster than even computing the empirical covariance matrix.

(0 min) Understanding the training dynamics of deep learning models is perhaps a necessary step toward demystifying the effectiveness of these models. In particular, how do data from different classes gradually become separable in their feature spaces when training neural networks using stochastic gradient descent? In this study, we model the evolution of features during deep learning training using a set of stochastic differential equations (SDEs) that each corresponds to a training sample. As a crucial ingredient in our modeling strategy, each SDE contains a drift term that reflects the impact of backpropagation at an input on the features of all samples. Our main finding uncovers a sharp phase transition phenomenon regarding the {intra-class impact: if the SDEs are locally elastic in the sense that the impact is more significant on samples from the same class as the input, the features of the training data become linearly separable, meaning vanishing training loss; otherwise, the features are not separable, regardless of how long the training time is. Moreover, in the presence of local elasticity, an analysis of our SDEs shows that the emergence of a simple geometric structure called the neural collapse of the features. Taken together, our results shed light on the decisive role of local elasticity in the training dynamics of neural networks. We corroborate our theoretical analysis with experiments on a synthesized dataset of geometric shapes and CIFAR-10.

(0 min) Real-time EEG-based Emotion Recognition (EEG-ER) with consumer-grade EEG devices involves classification of emotions using a reduced number of channels. These devices typically provide only four or five channels, unlike the high number of channels (32 or more) typically used in most current state-of-the-art research. In this work we propose to use Discrete Wavelet Transforms (DWT) to extract time-frequency domain features, and we use time-windows of a few seconds to perform EEG-ER classification. This technique can be used in real-time, as opposed to post-hoc on the full session data. We also apply baseline removal preprocessing, developed in prior research, to our proposed DWT Entropy and Energy features, which improves classification accuracy significantly. We consider two different classifier architectures, a 3D Convolutional Neural Network (3D CNN) and a Support Vector Machine (SVM). We evaluate both models on subject-independent and subject dependent setups to classify the Valence and Arousal dimensions of an individual's emotional state. We test them on both the full 32-channel data provided by the DEAP dataset, and also a reduced 5-channel extract of the same dataset. The SVM model performs best on all the presented scenarios, achieving an accuracy of 95.32% on Valence and 95.68% on Arousal for the full 32-channel subject-dependent case, beating prior real-time EEG-ER subject-dependent benchmarks. On the subject-independent case an accuracy of 80.70% on Valence and 81.41% on Arousal was also obtained. Reducing the input data to 5 channels only degrades the accuracy by an average of 3.54% across all scenarios, making this model appropriate for use with more accessible low-end EEG devices.

(0 min) Poisoning attacks on machine learning systems compromise the model performance by deliberately injecting malicious samples in the training dataset to influence the training process. Prior works focus on either availability attacks (i.e., lowering the overall model accuracy) or integrity attacks (i.e., enabling specific instance-based backdoor). In this paper, we advance the adversarial objectives of the availability attacks to a per-class basis, which we refer to as class-oriented poisoning attacks. We demonstrate that the proposed attack is capable of forcing the corrupted model to predict in two specific ways: (i) classify unseen new images to a targeted "supplanter" class, and (ii) misclassify images from a "victim" class while maintaining the classification accuracy on other non-victim classes. To maximize the adversarial effect as well as reduce the computational complexity of poisoned data generation, we propose a gradient-based framework that crafts poisoning images with carefully manipulated feature information for each scenario. Using newly defined metrics at the class level, we demonstrate the effectiveness of the proposed class-oriented poisoning attacks on various models (e.g., LeNet-5, Vgg-9, and ResNet-50) over a wide range of datasets (e.g., MNIST, CIFAR-10, and ImageNet-ILSVRC2012) in an end-to-end training setting.

(0 min) This paper aims to address the major challenges of Federated Learning (FL) on edge devices: limited memory and expensive communication. We propose a novel method, called Partial Variable Training (PVT), that only trains a small subset of variables on edge devices to reduce memory usage and communication cost. With PVT, we show that network accuracy can be maintained by utilizing more local training steps and devices, which is favorable for FL involving a large population of devices. According to our experiments on two state-of-the-art neural networks for speech recognition and two different datasets, PVT can reduce memory usage by up to 1.9$\times$ and communication cost by up to 593$\times$ while attaining comparable accuracy when compared with full network training.

(0 min) Recent work suggests that quantum machine learning techniques can be used for classical image classification by encoding the images in quantum states and using a quantum neural network for inference. However, such work has been restricted to very small input images, at most 4 x 4, that are unrealistic and cannot even be accurately labeled by humans. The primary difficulties in using larger input images is that hitherto-proposed encoding schemes necessitate more qubits than are physically realizable. We propose a framework to classify larger, realistic images using quantum systems. Our approach relies on a novel encoding mechanism that embeds images in quantum states while necessitating fewer qubits than prior work. Our framework is able to classify images that are larger than previously possible, up to 16 x 16 for the MNIST dataset on a personal laptop, and obtains accuracy comparable to classical neural networks with the same number of learnable parameters. We also propose a technique for further reducing the number of qubits needed to represent images that may result in an easier physical implementation at the expense of final performance. Our work enables quantum machine learning and classification on classical datasets of dimensions that were previously intractable by physically realizable quantum computers or classical simulation

(0 min) Deep neural networks suffer from poor generalization to unseen environments when the underlying data distribution is different from that in the training set. By learning minimum sufficient representations from training data, the information bottleneck (IB) approach has demonstrated its effectiveness to improve generalization in different AI applications. In this work, we propose a new neural network-based IB approach, termed gated information bottleneck (GIB), that dynamically drops spurious correlations and progressively selects the most task-relevant features across different environments by a trainable soft mask (on raw features). GIB enjoys a simple and tractable objective, without any variational approximation or distributional assumption. We empirically demonstrate the superiority of GIB over other popular neural network-based IB approaches in adversarial robustness and out-of-distribution (OOD) detection. Meanwhile, we also establish the connection between IB theory and invariant causal representation learning, and observed that GIB demonstrates appealing performance when different environments arrive sequentially, a more practical scenario where invariant risk minimization (IRM) fails. Code of GIB is available at https://github.com/falesiani/GIB

(0 min) We describe an implicit sparsity-inducing mechanism based on minimization over a family of kernels: \begin{equation*} \min_{\beta, f}~\widehat{\mathbb{E}}[L(Y, f(\beta^{1/q} \odot X)] + \lambda_n \|f\|_{\mathcal{H}_q}^2~~\text{subject to}~~\beta \ge 0, \end{equation*} where $L$ is the loss, $\odot$ is coordinate-wise multiplication and $\mathcal{H}_q$ is the reproducing kernel Hilbert space based on the kernel $k_q(x, x') = h(\|x-x'\|_q^q)$, where $\|\cdot\|_q$ is the $\ell_q$ norm. Using gradient descent to optimize this objective with respect to $\beta$ leads to exactly sparse stationary points with high probability. The sparsity is achieved without using any of the well-known explicit sparsification techniques such as penalization (e.g., $\ell_1$), early stopping or post-processing (e.g., clipping). As an application, we use this sparsity-inducing mechanism to build algorithms consistent for feature selection.

(0 min) Dermatologists often diagnose or rule out early melanoma by evaluating the follow-up dermoscopic images of skin lesions. However, existing algorithms for early melanoma diagnosis are developed using single time-point images of lesions. Ignoring the temporal, morphological changes of lesions can lead to misdiagnosis in borderline cases. In this study, we propose a framework for automated early melanoma diagnosis using sequential dermoscopic images. To this end, we construct our method in three steps. First, we align sequential dermoscopic images of skin lesions using estimated Euclidean transformations, extract the lesion growth region by computing image differences among the consecutive images, and then propose a spatio-temporal network to capture the dermoscopic changes from aligned lesion images and the corresponding difference images. Finally, we develop an early diagnosis module to compute probability scores of malignancy for lesion images over time. We collected 179 serial dermoscopic imaging data from 122 patients to verify our method. Extensive experiments show that the proposed model outperforms other commonly used sequence models. We also compared the diagnostic results of our model with those of seven experienced dermatologists and five registrars. Our model achieved higher diagnostic accuracy than clinicians (63.69% vs. 54.33%, respectively) and provided an earlier diagnosis of melanoma (60.7% vs. 32.7% of melanoma correctly diagnosed on the first follow-up images). These results demonstrate that our model can be used to identify melanocytic lesions that are at high-risk of malignant transformation earlier in the disease process and thereby redefine what is possible in the early detection of melanoma.

(0 min) Autonomous driving is an active research topic in both academia and industry. However, most of the existing solutions focus on improving the accuracy by training learnable models with centralized large-scale data. Therefore, these methods do not take into account the user's privacy. In this paper, we present a new approach to learn autonomous driving policy while respecting privacy concerns. We propose a peer-to-peer Deep Federated Learning (DFL) approach to train deep architectures in a fully decentralized manner and remove the need for central orchestration. We design a new Federated Autonomous Driving network (FADNet) that can improve the model stability, ensure convergence, and handle imbalanced data distribution problems while is being trained with federated learning methods. Intensively experimental results on three datasets show that our approach with FADNet and DFL achieves superior accuracy compared with other recent methods. Furthermore, our approach can maintain privacy by not collecting user data to a central server.

(0 min) Adequate labeled data and expensive compute resources are the prerequisites for the success of neural architecture search(NAS). It is challenging to apply NAS in meta-learning scenarios with limited compute resources and data. In this paper, an across-task neural architecture search (AT-NAS) is proposed to address the problem through combining gradient-based meta-learning with EA-based NAS to learn over the distribution of tasks. The supernet is learned over an entire set of tasks by meta-learning its weights. Architecture encodes of subnets sampled from the supernet are iteratively adapted by evolutionary algorithms while simultaneously searching for a task-sensitive meta-network. Searched meta-network can be adapted to a novel task via a few learning steps and only costs a little search time. Empirical results show that AT-NAS surpasses the related approaches on few-shot classification accuracy. The performance of AT-NAS on classification benchmarks is comparable to that of models searched from scratch, by adapting the architecture in less than an hour from a 5-GPU-day pretrained meta-network.

(0 min) The recently proposed Mean Teacher has achieved state-of-the-art results in several semi-supervised learning benchmarks. The Mean Teacher method can exploit large-scale unlabeled data in a self-ensembling manner. In this paper, an effective Couple Learning method based on a well-trained model and a Mean Teacher model is proposed. The proposed pseudo-labels generated model (PLG) can increase strongly-labeled data and weakly-labeled data to improve performance of the Mean Teacher method. The Mean Teacher method can suppress noise in pseudo-labels data. The Couple Learning method can extract more information in the compound training data. These experimental results on Task 4 of the DCASE2020 challenge demonstrate the superiority of the proposed method, achieving about 39.18% F1-score on public eval set, outperforming 37.12% of the baseline system by a significant margin.

(0 min) Generating the periodic structure of stable materials is a long-standing challenge for the material design community. This task is difficult because stable materials only exist in a low-dimensional subspace of all possible periodic arrangements of atoms: 1) the coordinates must lie in the local energy minimum defined by quantum mechanics, and 2) global stability also requires the structure to follow the complex, yet specific bonding preferences between different atom types. Existing methods fail to incorporate these factors and often lack proper invariances. We propose a Crystal Diffusion Variational Autoencoder (CDVAE) that captures the physical inductive bias of material stability. By learning from the data distribution of stable materials, the decoder generates materials in a diffusion process that moves atomic coordinates towards a lower energy state and updates atom types to satisfy bonding preferences between neighbors. Our model also explicitly encodes interactions across periodic boundaries and respects permutation, translation, rotation, and periodic invariances. We significantly outperform past methods in three tasks: 1) reconstructing the input structure, 2) generating valid, diverse, and realistic materials, and 3) generating materials that optimize a specific property. We also provide several standard datasets and evaluation metrics for the broader machine learning community.

(0 min) Mental disorders are among the leading causes of disability worldwide. The first step in treating these conditions is to obtain an accurate diagnosis, but the absence of established clinical tests makes this task challenging. Machine learning algorithms can provide a possible solution to this problem, as we describe in this work. We present a method for the automatic diagnosis of mental disorders based on the matrix of connections obtained from EEG time series and deep learning. We show that our approach can classify patients with Alzheimer's disease and schizophrenia with a high level of accuracy. The comparison with the traditional cases, that use raw EEG time series, shows that our method provides the highest precision. Therefore, the application of deep neural networks on data from brain connections is a very promising method to the diagnosis of neurological disorders.

(0 min) We present temporally abstract actor-critic (TAAC), a simple but effective off-policy RL algorithm that incorporates closed-loop temporal abstraction into the actor-critic framework. TAAC adds a second-stage binary policy to choose between the previous action and a new action output by an actor. Crucially, its "act-or-repeat" decision hinges on the actually sampled action instead of the expected behavior of the actor. This post-acting switching scheme let the overall policy make more informed decisions. TAAC has two important features: a) persistent exploration, and b) a new compare-through Q operator for multi-step TD backup, specially tailored to the action repetition scenario. We demonstrate TAAC's advantages over several strong baselines across 14 continuous control tasks. Our surprising finding reveals that while achieving top performance, TAAC is able to "mine" a significant number of repeated actions with the trained policy even on continuous tasks whose problem structures on the surface seem to repel action repetition. This suggests that aside from encouraging persistent exploration, action repetition can find its place in a good policy behavior. Code is available at https://github.com/hnyu/taac.

(0 min) Human gait can be a predictive factor for detecting pathologies that affect human locomotion according to studies. In addition, it is known that a high investment is demanded in order to raise a traditional clinical infrastructure able to provide human gait examinations, making them unaffordable for economically vulnerable patients. In face of this scenario, this work proposes an accessible and modern solution composed of a wearable device, to acquire 3D-accelerometer and 3D-gyroscope measurements, and machine learning techniques to classify between distinct categories of induced gait disorders. In order to develop the proposed research, it was created a dataset with the target label being 4 distinct and balanced categories of anomalous gait. The machine learning techniques that achieved the best performances (in terms of accuracy) in this dataset were through the application of Principal Component Analysis algorithm following of a Support Vector Machines classifier (94 \%). Further, an architecture based on a Feedforward Neural Network yielded even better results (96 \%). Finally, it is also presented computational performance comparison between the models implemented.

(0 min) R\'enyi's information provides a theoretical foundation for tractable and data-efficient non-parametric density estimation, based on pair-wise evaluations in a reproducing kernel Hilbert space (RKHS). This paper extends this framework to parametric probabilistic modeling, motivated by the fact that R\'enyi's information can be estimated in closed-form for Gaussian mixtures. Based on this special connection, a novel generative model framework called the structured generative model (SGM) is proposed that makes straightforward optimization possible, because costs are scale-invariant, avoiding high gradient variance while imposing less restrictions on absolute continuity, which is a huge advantage in parametric information theoretic optimization. The implementation employs a single neural network driven by an orthonormal input appended to a single white noise source adapted to learn an infinite Gaussian mixture model (IMoG), which provides an empirically tractable model distribution in low dimensions. To train SGM, we provide three novel variational cost functions, based on R\'enyi's second-order entropy and divergence, to implement minimization of cross-entropy, minimization of variational representations of $f$-divergence, and maximization of the evidence lower bound (conditional probability). We test the framework for estimation of mutual information and compare the results with the mutual information neural estimation (MINE), for density estimation, for conditional probability estimation in Markov models as well as for training adversarial networks. Our preliminary results show that SGM significantly improves MINE estimation in terms of data efficiency and variance, conventional and variational Gaussian mixture models, as well as the performance of generative adversarial networks.

(0 min) Most existing neural architecture search (NAS) benchmarks and algorithms prioritize performance on well-studied tasks, e.g., image classification on CIFAR and ImageNet. This makes the applicability of NAS approaches in more diverse areas inadequately understood. In this paper, we present NAS-Bench-360, a benchmark suite for evaluating state-of-the-art NAS methods for convolutional neural networks (CNNs). To construct it, we curate a collection of ten tasks spanning a diverse array of application domains, dataset sizes, problem dimensionalities, and learning objectives. By carefully selecting tasks that can both interoperate with modern CNN-based search methods but that are also far-afield from their original development domain, we can use NAS-Bench-360 to investigate the following central question: do existing state-of-the-art NAS methods perform well on diverse tasks? Our experiments show that a modern NAS procedure designed for image classification can indeed find good architectures for tasks with other dimensionalities and learning objectives; however, the same method struggles against more task-specific methods and performs catastrophically poorly on classification in non-vision domains. The case for NAS robustness becomes even more dire in a resource-constrained setting, where a recent NAS method provides little-to-no benefit over much simpler baselines. These results demonstrate the need for a benchmark such as NAS-Bench-360 to help develop NAS approaches that work well on a variety of tasks, a crucial component of a truly robust and automated pipeline. We conclude with a demonstration of the kind of future research our suite of tasks will enable. All data and code is made publicly available.

(0 min) There is currently a large interest in understanding the potential advantages quantum devices can offer for probabilistic modelling. In this work we investigate, within two different oracle models, the probably approximately correct (PAC) learnability of quantum circuit Born machines, i.e., the output distributions of local quantum circuits. We first show a negative result, namely, that the output distributions of super-logarithmic depth Clifford circuits are not sample-efficiently learnable in the statistical query model, i.e., when given query access to empirical expectation values of bounded functions over the sample space. This immediately implies the hardness, for both quantum and classical algorithms, of learning from statistical queries the output distributions of local quantum circuits using any gate set which includes the Clifford group. As many practical generative modelling algorithms use statistical queries -- including those for training quantum circuit Born machines -- our result is broadly applicable and strongly limits the possibility of a meaningful quantum advantage for learning the output distributions of local quantum circuits. As a positive result, we show that in a more powerful oracle model, namely when directly given access to samples, the output distributions of local Clifford circuits are computationally efficiently PAC learnable by a classical learner. Our results are equally applicable to the problems of learning an algorithm for generating samples from the target distribution (generative modelling) and learning an algorithm for evaluating its probabilities (density modelling). They provide the first rigorous insights into the learnability of output distributions of local quantum circuits from the probabilistic modelling perspective.

(0 min) We observe that despite their hierarchical convolutional nature, the synthesis process of typical generative adversarial networks depends on absolute pixel coordinates in an unhealthy manner. This manifests itself as, e.g., detail appearing to be glued to image coordinates instead of the surfaces of depicted objects. We trace the root cause to careless signal processing that causes aliasing in the generator network. Interpreting all signals in the network as continuous, we derive generally applicable, small architectural changes that guarantee that unwanted information cannot leak into the hierarchical synthesis process. The resulting networks match the FID of StyleGAN2 but differ dramatically in their internal representations, and they are fully equivariant to translation and rotation even at subpixel scales. Our results pave the way for generative models better suited for video and animation.

(0 min) Multi-agent reinforcement learning (MARL) has attracted much research attention recently. However, unlike its single-agent counterpart, many theoretical and algorithmic aspects of MARL have not been well-understood. In this paper, we study the emergence of coordinated behavior by autonomous agents using an actor-critic (AC) algorithm. Specifically, we propose and analyze a class of coordinated actor-critic algorithms (CAC) in which individually parametrized policies have a {\it shared} part (which is jointly optimized among all agents) and a {\it personalized} part (which is only locally optimized). Such kind of {\it partially personalized} policy allows agents to learn to coordinate by leveraging peers' past experience and adapt to individual tasks. The flexibility in our design allows the proposed MARL-CAC algorithm to be used in a {\it fully decentralized} setting, where the agents can only communicate with their neighbors, as well as a {\it federated} setting, where the agents occasionally communicate with a server while optimizing their (partially personalized) local models. Theoretically, we show that under some standard regularity assumptions, the proposed MARL-CAC algorithm requires $\mathcal{O}(\epsilon^{-\frac{5}{2}})$ samples to achieve an $\epsilon$-stationary solution (defined as the solution whose squared norm of the gradient of the objective function is less than $\epsilon$). To the best of our knowledge, this work provides the first finite-sample guarantee for decentralized AC algorithm with partially personalized policies.

(0 min) One of the main reasons for Alzheimer's disease (AD) is the disorder of some neural circuits. Existing methods for AD prediction have achieved great success, however, detecting abnormal neural circuits from the perspective of brain networks is still a big challenge. In this work, a novel decoupling generative adversarial network (DecGAN) is proposed to detect abnormal neural circuits for AD. Concretely, a decoupling module is designed to decompose a brain network into two parts: one part is composed of a few sparse graphs which represent the neural circuits largely determining the development of AD; the other part is a supplement graph, whose influence on AD can be ignored. Furthermore, the adversarial strategy is utilized to guide the decoupling module to extract the feature more related to AD. Meanwhile, by encoding the detected neural circuits to hypergraph data, an analytic module associated with the hyperedge neurons algorithm is designed to identify the neural circuits. More importantly, a novel sparse capacity loss based on the spatial-spectral hypergraph similarity is developed to minimize the intrinsic topological distribution of neural circuits, which can significantly improve the accuracy and robustness of the proposed model. Experimental results demonstrate that the proposed model can effectively detect the abnormal neural circuits at different stages of AD, which is helpful for pathological study and early treatment.

(0 min) Understanding the fundamental mechanism behind the success of deep neural networks is one of the key challenges in the modern machine learning literature. Despite numerous attempts, a solid theoretical analysis is yet to be developed. In this paper, we develop a novel unified framework to reveal a hidden regularization mechanism through the lens of convex optimization. We first show that the training of multiple three-layer ReLU sub-networks with weight decay regularization can be equivalently cast as a convex optimization problem in a higher dimensional space, where sparsity is enforced via a group $\ell_1$-norm regularization. Consequently, ReLU networks can be interpreted as high dimensional feature selection methods. More importantly, we then prove that the equivalent convex problem can be globally optimized by a standard convex optimization solver with a polynomial-time complexity with respect to the number of samples and data dimension when the width of the network is fixed. Finally, we numerically validate our theoretical results via experiments involving both synthetic and real datasets.

(0 min) We propose a new line-search method, coined Curvature-Aware Random Search (CARS), for derivative-free optimization. CARS exploits approximate curvature information to estimate the optimal step-size given a search direction. We prove that for strongly convex objective functions, CARS converges linearly if the search direction is drawn from a distribution satisfying very mild conditions. We also explore a variant, CARS-NQ, which uses Numerical Quadrature instead of a Monte Carlo method when approximating curvature along the search direction. We show CARS-NQ is effective on highly non-convex problems of the form $f = f_{\mathrm{cvx}} + f_{\mathrm{osc}}$ where $f_{\mathrm{cvx}}$ is strongly convex and $f_{\mathrm{osc}}$ is rapidly oscillating. Experimental results show that CARS and CARS-NQ match or exceed the state-of-the-arts on benchmark problem sets.

(0 min) High-quality data accumulation is now becoming ubiquitous in the health domain. There is increasing opportunity to exploit rich data from normal subjects to improve supervised estimators in specific diseases with notorious data scarcity. We demonstrate that low-dimensional embedding spaces can be derived from the UK Biobank population dataset and used to enhance data-scarce prediction of health indicators, lifestyle and demographic characteristics. Phenotype predictions facilitated by Variational Autoencoder manifolds typically scaled better with increasing unlabeled data than dimensionality reduction by PCA or Isomap. Performances gains from semisupervison approaches will probably become an important ingredient for various medical data science applications.

(0 min) Continuous speech separation using a microphone array was shown to be promising in dealing with the speech overlap problem in natural conversation transcription. This paper proposes VarArray, an array-geometry-agnostic speech separation neural network model. The proposed model is applicable to any number of microphones without retraining while leveraging the nonlinear correlation between the input channels. The proposed method adapts different elements that were proposed before separately, including transform-average-concatenate, conformer speech separation, and inter-channel phase differences, and combines them in an efficient and cohesive way. Large-scale evaluation was performed with two real meeting transcription tasks by using a fully developed transcription system requiring no prior knowledge such as reference segmentations, which allowed us to measure the impact that the continuous speech separation system could have in realistic settings. The proposed model outperformed a previous approach to array-geometry-agnostic modeling for all of the geometry configurations considered, achieving asclite-based speaker-agnostic word error rates of 17.5% and 20.4% for the AMI development and evaluation sets, respectively, in the end-to-end setting using no ground-truth segmentations.

(0 min) Experiments to understand the sensorimotor neural interactions in the human cortical speech system support the existence of a bidirectional flow of interactions between the auditory and motor regions. Their key function is to enable the brain to 'learn' how to control the vocal tract for speech production. This idea is the impetus for the recently proposed "MirrorNet", a constrained autoencoder architecture. In this paper, the MirrorNet is applied to learn, in an unsupervised manner, the controls of a specific audio synthesizer (DIVA) to produce melodies only from their auditory spectrograms. The results demonstrate how the MirrorNet discovers the synthesizer parameters to generate the melodies that closely resemble the original and those of unseen melodies, and even determine the best set parameters to approximate renditions of complex piano melodies generated by a different synthesizer. This generalizability of the MirrorNet illustrates its potential to discover from sensory data the controls of arbitrary motor-plants such as autonomous vehicles.

(2 min) Data augmentation methods have shown great importance in diverse supervised learning problems where labeled data is scarce or costly to obtain. For sound event localization and detection (SELD) tasks several augmentation methods have been proposed, with most borrowing ideas from other domains such as images, speech, or monophonic audio. However, only a few exploit the spatial properties of a full 3D audio scene. We propose Spatial Mixup, as an application of parametric spatial audio effects for data augmentation, which modifies the directional properties of a multi-channel spatial audio signal encoded in the ambisonics domain. Similarly to beamforming, these modifications enhance or suppress signals arriving from certain directions, although the effect is less pronounced. Therefore enabling deep learning models to achieve invariance to small spatial perturbations. The method is evaluated with experiments in the DCASE 2021 Task 3 dataset, where spatial mixup increases performance over a non-augmented baseline, and compares to other well known augmentation methods. Furthermore, combining spatial mixup with other methods greatly improves performance.

(2 min) Understanding and explaining the mistakes made by trained models is critical to many machine learning objectives, such as improving robustness, addressing concept drift, and mitigating biases. However, this is often an ad hoc process that involves manually looking at the model's mistakes on many test samples and guessing at the underlying reasons for those incorrect predictions. In this paper, we propose a systematic approach, conceptual counterfactual explanations(CCE), that explains why a classifier makes a mistake on a particular test sample(s) in terms of human-understandable concepts (e.g. this zebra is misclassified as a dog because of faint stripes). We base CCE on two prior ideas: counterfactual explanations and concept activation vectors, and validate our approach on well-known pretrained models, showing that it explains the models' mistakes meaningfully. In addition, for new models trained on data with spurious correlations, CCE accurately identifies the spurious correlation as the cause of model mistakes from a single misclassified test sample. On two challenging medical applications, CCE generated useful insights, confirmed by clinicians, into biases and mistakes the model makes in real-world settings. The code for CCE is publicly available and can easily be applied to explain mistakes in new models.

(2 min) We present a method for conditional sampling for pre-trained normalizing flows when only part of an observation is available. We derive a lower bound to the conditioning variable log-probability using Schur complement properties in the spirit of Gaussian conditional sampling. Our derivation relies on partitioning flow's domain in such a way that the flow restrictions to subdomains remain bijective, which is crucial for the Schur complement application. Simulation from the variational conditional flow then amends to solving an equality constraint. Our contribution is three-fold: a) we provide detailed insights on the choice of variational distributions; b) we discuss how to partition the input space of the flow to preserve bijectivity property; c) we propose a set of methods to optimise the variational distribution. Our numerical results indicate that our sampling method can be successfully applied to invertible residual networks for inference and classification.

(0 min) Modeling biological dynamical systems is challenging due to the interdependence of different system components, some of which are not fully understood. To fill existing gaps in our ability to mechanistically model physiological systems, we propose to combine neural networks with physics-based models. Specifically, we demonstrate how we can approximate missing ordinary differential equations (ODEs) coupled with known ODEs using Bayesian filtering techniques to train the model parameters and simultaneously estimate dynamic state variables. As a study case we leverage a well-understood model for blood circulation in the human retina and replace one of its core ODEs with a neural network approximation, representing the case where we have incomplete knowledge of the physiological state dynamics. Results demonstrate that state dynamics corresponding to the missing ODEs can be approximated well using a neural network trained using a recursive Bayesian filtering approach in a fashion coupled with the known state dynamic differential equations. This demonstrates that dynamics and impact of missing state variables can be captured through joint state estimation and model parameter estimation within a recursive Bayesian state estimation (RBSE) framework. Results also indicate that this RBSE approach to training the NN parameters yields better outcomes (measurement/state estimation accuracy) than training the neural network with backpropagation through time in the same setting.

(0 min) Vision Transformers (ViTs) and MLPs signal further efforts on replacing hand-wired features or inductive biases with general-purpose neural architectures. Existing works empower the models by massive data, such as large-scale pre-training and/or repeated strong data augmentations, and still report optimization-related problems (e.g., sensitivity to initialization and learning rates). Hence, this paper investigates ViTs and MLP-Mixers from the lens of loss geometry, intending to improve the models' data efficiency at training and generalization at inference. Visualization and Hessian reveal extremely sharp local minima of converged models. By promoting smoothness with a recently proposed sharpness-aware optimizer, we substantially improve the accuracy and robustness of ViTs and MLP-Mixers on various tasks spanning supervised, adversarial, contrastive, and transfer learning (e.g., +5.3\% and +11.0\% top-1 accuracy on ImageNet for ViT-B/16 and Mixer-B/16, respectively, with the simple Inception-style preprocessing). We show that the improved smoothness attributes to sparser active neurons in the first few layers. The resultant ViTs outperform ResNets of similar size and throughput when trained from scratch on ImageNet without large-scale pre-training or strong data augmentations. They also possess more perceptive attention maps. Our model checkpoints are released at \url{https://github.com/google-research/vision_transformer}.

(2 min) In this paper, we investigate the performance of two first-order optimization algorithms, obtained from forward Euler discretization of finite-time optimization flows. These flows are the rescaled-gradient flow (RGF) and the signed-gradient flow (SGF), and consist of non-Lipscthiz or discontinuous dynamical systems that converge locally in finite time to the minima of gradient-dominated functions. We propose an Euler discretization for these first-order finite-time flows, and provide convergence guarantees, in the deterministic and the stochastic setting. We then apply the proposed algorithms to academic examples, as well as deep neural networks training, where we empirically test their performances on the SVHN dataset. Our results show that our schemes demonstrate faster convergences against standard optimization alternatives.

(2 min) One of the central puzzles in modern machine learning is the ability of heavily overparametrized models to generalize well. Although the low-dimensional structure of typical datasets is key to this behavior, most theoretical studies of overparametrization focus on isotropic inputs. In this work, we instead consider an analytically tractable model of structured data, where the input covariance is built from independent blocks allowing us to tune the saliency of low-dimensional structures and their alignment with respect to the target function. Using methods from statistical physics, we derive a precise asymptotic expression for the train and test error achieved by random feature models trained to classify such data, which is valid for any convex loss function. We study in detail how the data structure affects the double descent curve, and show that in the over-parametrized regime, its impact is greater for logistic loss than for mean-squared loss: the easier the task, the wider the gap in performance at the advantage of the logistic loss. Our insights are confirmed by numerical experiments on MNIST and CIFAR10.

(2 min) To achieve systematic generalisation, it first makes sense to master simple tasks such as arithmetic. Of the four fundamental arithmetic operations (+,-,$\times$,$\div$), division is considered the most difficult for both humans and computers. In this paper we show that robustly learning division in a systematic manner remains a challenge even at the simplest level of dividing two numbers. We propose two novel approaches for division which we call the Neural Reciprocal Unit (NRU) and the Neural Multiplicative Reciprocal Unit (NMRU), and present improvements for an existing division module, the Real Neural Power Unit (Real NPU). Experiments in learning division with input redundancy on 225 different training sets, find that our proposed modifications to the Real NPU obtains an average success of 85.3$\%$ improving over the original by 15.1$\%$. In light of the suggestion above, our NMRU approach can further improve the success to 91.6$\%$.

(0 min) Anomaly detection is an active research topic in many different fields such as intrusion detection, network monitoring, system health monitoring, IoT healthcare, etc. However, many existing anomaly detection approaches require either human intervention or domain knowledge, and may suffer from high computation complexity, consequently hindering their applicability in real-world scenarios. Therefore, a lightweight and ready-to-go approach that is able to detect anomalies in real-time is highly sought-after. Such an approach could be easily and immediately applied to perform time series anomaly detection on any commodity machine. The approach could provide timely anomaly alerts and by that enable appropriate countermeasures to be undertaken as early as possible. With these goals in mind, this paper introduces ReRe, which is a Real-time Ready-to-go proactive Anomaly Detection algorithm for streaming time series. ReRe employs two lightweight Long Short-Term Memory (LSTM) models to predict and jointly determine whether or not an upcoming data point is anomalous based on short-term historical data points and two long-term self-adaptive thresholds. Experiments based on real-world time-series datasets demonstrate the good performance of ReRe in real-time anomaly detection without requiring human intervention or domain knowledge.

(0 min) Modern deep neural networks are highly over-parameterized compared to the data on which they are trained, yet they often generalize remarkably well. A flurry of recent work has asked: why do deep networks not overfit to their training data? In this work, we make a series of empirical observations that investigate the hypothesis that deeper networks are inductively biased to find solutions with lower rank embeddings. We conjecture that this bias exists because the volume of functions that maps to low-rank embedding increases with depth. We show empirically that our claim holds true on finite width linear and non-linear models and show that these are the solutions that generalize well. We then show that the low-rank simplicity bias exists even after training, using a wide variety of commonly used optimizers. We found this phenomenon to be resilient to initialization, hyper-parameters, and learning methods. We further demonstrate how linear over-parameterization of deep non-linear models can be used to induce low-rank bias, improving generalization performance without changing the effective model capacity. Practically, we demonstrate that simply linearly over-parameterizing standard models at training time can improve performance on image classification tasks, including ImageNet.

(0 min) Inspired by diversity of biological neurons, quadratic artificial neurons can play an important role in deep learning models. The type of quadratic neurons of our interest replaces the inner-product operation in the conventional neuron with a quadratic function. Despite promising results so far achieved by networks of quadratic neurons, there are important issues not well addressed. Theoretically, the superior expressivity of a quadratic network over either a conventional network or a conventional network via quadratic activation is not fully elucidated, which makes the use of quadratic networks not well grounded. Practically, although a quadratic network can be trained via generic backpropagation, it can be subject to a higher risk of collapse than the conventional counterpart. To address these issues, we first apply the spline theory and a measure from algebraic geometry to give two theorems that demonstrate better model expressivity of a quadratic network than the conventional counterpart with or without quadratic activation. Then, we propose an effective and efficient training strategy referred to as ReLinear to stabilize the training process of a quadratic network, thereby unleashing the full potential in its associated machine learning tasks. Comprehensive experiments on popular datasets are performed to support our findings and evaluate the performance of quadratic deep learning.

(0 min) Event Extraction bridges the gap between text and event signals. Based on the assumption of trigger-argument dependency, existing approaches have achieved state-of-the-art performance with expert-designed templates or complicated decoding constraints. In this paper, for the first time we introduce the prompt-based learning strategy to the domain of Event Extraction, which empowers the automatic exploitation of label semantics on both input and output sides. To validate the effectiveness of the proposed generative method, we conduct extensive experiments with 11 diverse baselines. Empirical results show that, in terms of F1 score on Argument Extraction, our simple architecture is stronger than any other generative counterpart and even competitive with algorithms that require template engineering. Regarding the measure of recall, it sets new overall records for both Argument and Trigger Extractions. We hereby recommend this framework to the community, with the code publicly available at https://git.io/GDAP.

(0 min) To enable robots to instruct humans in collaborations, we identify several aspects of language processing that are not commonly studied in this context. These include location, planning, and generation. We suggest evaluations for each task, offer baselines for simple methods, and close by discussing challenges and opportunities in studying language for collaboration.

(0 min) The problem of answering questions using knowledge from pre-trained language models (LMs) and knowledge graphs (KGs) presents two challenges: given a QA context (question and answer choice), methods need to (i) identify relevant knowledge from large KGs, and (ii) perform joint reasoning over the QA context and KG. In this work, we propose a new model, QA-GNN, which addresses the above challenges through two key innovations: (i) relevance scoring, where we use LMs to estimate the importance of KG nodes relative to the given QA context, and (ii) joint reasoning, where we connect the QA context and KG to form a joint graph, and mutually update their representations through graph neural networks. We evaluate QA-GNN on the CommonsenseQA and OpenBookQA datasets, and show its improvement over existing LM and LM+KG models, as well as its capability to perform interpretable and structured reasoning, e.g., correctly handling negation in questions.

(0 min) Online platforms and communities establish their own norms that govern what behavior is acceptable within the community. Substantial effort in NLP has focused on identifying unacceptable behaviors and, recently, on forecasting them before they occur. However, these efforts have largely focused on toxicity as the sole form of community norm violation. Such focus has overlooked the much larger set of rules that moderators enforce. Here, we introduce a new dataset focusing on a more complete spectrum of community norms and their violations in the local conversational and global community contexts. We introduce a series of models that use this data to develop context- and community-sensitive norm violation detection, showing that these changes give high performance.

(2 min) Recently proposed self-supervised learning approaches have been successful for pre-training speech representation models. The utility of these learned representations has been observed empirically, but not much has been studied about the type or extent of information encoded in the pre-trained representations themselves. Developing such insights can help understand the capabilities and limits of these models and enable the research community to more efficiently develop their usage for downstream applications. In this work, we begin to fill this gap by examining one recent and successful pre-trained model (wav2vec 2.0), via its intermediate representation vectors, using a suite of analysis tools. We use the metrics of canonical correlation, mutual information, and performance on simple downstream tasks with non-parametric probes, in order to (i) query for acoustic and linguistic information content, (ii) characterize the evolution of information across model layers, and (iii) understand how fine-tuning the model for automatic speech recognition (ASR) affects these observations. Our findings motivate modifying the fine-tuning protocol for ASR, which produces improved word error rates in a low-resource setting.

(2 min) Typically, information extraction (IE) requires a pipeline approach: first, a sequence labeling model is trained on manually annotated documents to extract relevant spans; then, when a new document arrives, a model predicts spans which are then post-processed and standardized to convert the information into a database entry. We replace this labor-intensive workflow with a transformer language model trained on existing database records to directly generate structured JSON. Our solution removes the workload associated with producing token-level annotations and takes advantage of a data source which is generally quite plentiful (e.g. database records). As long documents are common in information extraction tasks, we use gradient checkpointing and chunked encoding to apply our method to sequences of up to 32,000 tokens on a single GPU. Our Doc2Dict approach is competitive with more complex, hand-engineered pipelines and offers a simple but effective baseline for document-level information extraction. We release our Doc2Dict model and code to reproduce our experiments and facilitate future work.

(2 min) Language is an interface to the outside world. In order for embodied agents to use it, language must be grounded in other, sensorimotor modalities. While there is an extended literature studying how machines can learn grounded language, the topic of how to learn spatio-temporal linguistic concepts is still largely uncharted. To make progress in this direction, we here introduce a novel spatio-temporal language grounding task where the goal is to learn the meaning of spatio-temporal descriptions of behavioral traces of an embodied agent. This is achieved by training a truth function that predicts if a description matches a given history of observations. The descriptions involve time-extended predicates in past and present tense as well as spatio-temporal references to objects in the scene. To study the role of architectural biases in this task, we train several models including multimodal Transformer architectures; the latter implement different attention computations between words and objects across space and time. We test models on two classes of generalization: 1) generalization to randomly held-out sentences; 2) generalization to grammar primitives. We observe that maintaining object identity in the attention computation of our Transformers is instrumental to achieving good performance on generalization overall, and that summarizing object traces in a single token has little influence on performance. We then discuss how this opens new perspectives for language-guided autonomous embodied agents. We also release our code under open-source license as well as pretrained models and datasets to encourage the wider community to build upon and extend our work in the future.

(2 min) Multimodal abstractive summarization (MAS) models that summarize videos (vision modality) and their corresponding transcripts (text modality) are able to extract the essential information from massive multimodal data on the Internet. Recently, large-scale generative pre-trained language models (GPLMs) have been shown to be effective in text generation tasks. However, existing MAS models cannot leverage GPLMs' powerful generation ability. To fill this research gap, we aim to study two research questions: 1) how to inject visual information into GPLMs without hurting their generation ability; and 2) where is the optimal place in GPLMs to inject the visual information? In this paper, we present a simple yet effective method to construct vision guided (VG) GPLMs for the MAS task using attention-based add-on layers to incorporate visual information while maintaining their original text generation ability. Results show that our best model significantly surpasses the prior state-of-the-art model by 5.7 ROUGE-1, 5.3 ROUGE-2, and 5.1 ROUGE-L scores on the How2 dataset, and our visual guidance method contributes 83.6% of the overall improvement. Furthermore, we conduct thorough ablation studies to analyze the effectiveness of various modality fusion methods and fusion locations.

(2 min) While multi-party conversations are often less structured than monologues and documents, they are implicitly organized by semantic level correlations across the interactive turns, and dialogue discourse analysis can be applied to predict the dependency structure and relations between the elementary discourse units, and provide feature-rich structural information for downstream tasks. However, the existing corpora with dialogue discourse annotation are collected from specific domains with limited sample sizes, rendering the performance of data-driven approaches poor on incoming dialogues without any domain adaptation. In this paper, we first introduce a Transformer-based parser, and assess its cross-domain performance. We next adopt three methods to gain domain integration from both data and language modeling perspectives to improve the generalization capability. Empirical results show that the neural parser can benefit from our proposed methods, and performs better on cross-domain dialogue samples.

(2 min) Multi-hop Question Answering (QA) is a challenging task because it requires precise reasoning with entity relations at every step towards the answer. The relations can be represented in terms of labels in knowledge graph (e.g., \textit{spouse}) or text in text corpus (e.g., \textit{they have been married for 26 years}). Existing models usually infer the answer by predicting the sequential relation path or aggregating the hidden graph features. The former is hard to optimize, and the latter lacks interpretability. In this paper, we propose TransferNet, an effective and transparent model for multi-hop QA, which supports both label and text relations in a unified framework. TransferNet jumps across entities at multiple steps. At each step, it attends to different parts of the question, computes activated scores for relations, and then transfer the previous entity scores along activated relations in a differentiable way. We carry out extensive experiments on three datasets and demonstrate that TransferNet surpasses the state-of-the-art models by a large margin. In particular, on MetaQA, it achieves 100\% accuracy in 2-hop and 3-hop questions. By qualitative analysis, we show that TransferNet has transparent and interpretable intermediate results.

(2 min) We consider the problem of constructing matched groups such that the resulting groups are statistically similar with respect to their average values for multiple covariates. This group-matching problem arises in many cases, including quasi-experimental and observational studies in which subjects or items are sampled from pre-existing groups, scenarios in which traditional pair-matching approaches may be inappropriate. We consider the case in which one is provided with an existing sample and iteratively eliminates samples so that the groups "match" according to arbitrary statistically-defined criteria. This problem is NP-hard. However, using artificial and real-world data sets, we show that heuristics implemented by the ldamatch package produce high-quality matches.

(2 min) Recent studies in big data analytics and natural language processing develop automatic techniques in analyzing sentiment in the social media information. In addition, the growing user base of social media and the high volume of posts also provide valuable sentiment information to predict the price fluctuation of the cryptocurrency. This research is directed to predicting the volatile price movement of cryptocurrency by analyzing the sentiment in social media and finding the correlation between them. While previous work has been developed to analyze sentiment in English social media posts, we propose a method to identify the sentiment of the Chinese social media posts from the most popular Chinese social media platform Sina-Weibo. We develop the pipeline to capture Weibo posts, describe the creation of the crypto-specific sentiment dictionary, and propose a long short-term memory (LSTM) based recurrent neural network along with the historical cryptocurrency price movement to predict the price trend for future time frames. The conducted experiments demonstrate the proposed approach outperforms the state of the art auto regressive based model by 18.5% in precision and 15.4% in recall.

(2 min) This work presents a lifelong learning approach to train a multilingual Text-To-Speech (TTS) system, where each language was seen as an individual task and was learned sequentially and continually. It does not require pooled data from all languages altogether, and thus alleviates the storage and computation burden. One of the challenges of lifelong learning methods is "catastrophic forgetting": in TTS scenario it means that model performance quickly degrades on previous languages when adapted to a new language. We approach this problem via a data-replay-based lifelong learning method. We formulate the replay process as a supervised learning problem, and propose a simple yet effective dual-sampler framework to tackle the heavily language-imbalanced training samples. Through objective and subjective evaluations, we show that this supervised learning formulation outperforms other gradient-based and regularization-based lifelong learning methods, achieving 43% Mel-Cepstral Distortion reduction compared to a fine-tuning baseline.

(2 min) Bayesian Active Learning has had significant impact to various NLP problems, but nevertheless it's application to text summarization has been explored very little. We introduce Bayesian Active Summarization (BAS), as a method of combining active learning methods with state-of-the-art summarization models. Our findings suggest that BAS achieves better and more robust performance, compared to random selection, particularly for small and very small data annotation budgets. Using BAS we showcase it is possible to leverage large summarization models to effectively solve real-world problems with very limited annotated data.

(2 min) Referring Image Segmentation (RIS) aims at segmenting the target object from an image referred by one given natural language expression. The diverse and flexible expressions as well as complex visual contents in the images raise the RIS model with higher demands for investigating fine-grained matching behaviors between words in expressions and objects presented in images. However, such matching behaviors are hard to be learned and captured when the visual cues of referents (i.e. referred objects) are insufficient, as the referents with weak visual cues tend to be easily confused by cluttered background at boundary or even overwhelmed by salient objects in the image. And the insufficient visual cues issue can not be handled by the cross-modal fusion mechanisms as done in previous work. In this paper, we tackle this problem from a novel perspective of enhancing the visual information for the referents by devising a Two-stage Visual cues enhancement Network (TV-Net), where a novel Retrieval and Enrichment Scheme (RES) and an Adaptive Multi-resolution feature Fusion (AMF) module are proposed. Through the two-stage enhancement, our proposed TV-Net enjoys better performances in learning fine-grained matching behaviors between the natural language expression and image, especially when the visual information of the referent is inadequate, thus produces better segmentation results. Extensive experiments are conducted to validate the effectiveness of the proposed method on the RIS task, with our proposed TV-Net surpassing the state-of-the-art approaches on four benchmark datasets.

(2 min) Existing abstractive summarization models lack explicit control mechanisms that would allow users to influence the stylistic features of the model outputs. This results in generating generic summaries that do not cater to the users needs or preferences. To address this issue we introduce HydraSum, a new summarization architecture that extends the single decoder framework of current models, e.g. BART, to a mixture-of-experts version consisting of multiple decoders. Our proposed model encourages each expert, i.e. decoder, to learn and generate stylistically-distinct summaries along dimensions such as abstractiveness, length, specificity, and others. At each time step, HydraSum employs a gating mechanism that decides the contribution of each individual decoder to the next token's output probability distribution. Through experiments on three summarization datasets (CNN, Newsroom, XSum), we demonstrate that this gating mechanism automatically learns to assign contrasting summary styles to different HydraSum decoders under the standard training objective without the need for additional supervision. We further show that a guided version of the training process can explicitly govern which summary style is partitioned between decoders, e.g. high abstractiveness vs. low abstractiveness or high specificity vs. low specificity, and also increase the stylistic-difference between individual decoders. Finally, our experiments demonstrate that our decoder framework is highly flexible: during inference, we can sample from individual decoders or mixtures of different subsets of the decoders to yield a diverse set of summaries and enforce single- and multi-style control over summary generation.

(2 min) When a software bug is reported, developers engage in a discussion to collaboratively resolve it. While the solution is likely formulated within the discussion, it is often buried in a large amount of text, making it difficult to comprehend, which delays its implementation. To expedite bug resolution, we propose generating a concise natural language description of the solution by synthesizing relevant content within the discussion, which encompasses both natural language and source code. Furthermore, to support generating an informative description during an ongoing discussion, we propose a secondary task of determining when sufficient context about the solution emerges in real-time. We construct a dataset for these tasks with a novel technique for obtaining noisy supervision from repository changes linked to bug reports. We establish baselines for generating solution descriptions, and develop a classifier which makes a prediction following each new utterance on whether or not the necessary context for performing generation is available. Through automated and human evaluation, we find these tasks to form an ideal testbed for complex reasoning in long, bimodal dialogue context.

(2 min) Research in sociology and linguistics shows that people use language not only to express their own identity but to understand the identity of others. Recent work established a connection between expression of identity and emoji usage on social media, through use of emoji skin tone modifiers. Motivated by that finding, this work asks if, as with language, readers are sensitive to such acts of self-expression and use them to understand the identity of authors. In behavioral experiments (n=488), where text and emoji content of social media posts were carefully controlled before being presented to participants, we find in the affirmative -- emoji are a salient signal of author identity. That signal is distinct from, and complementary to, the one encoded in language. Participant groups (based on self-identified ethnicity) showed no differences in how they perceive this signal, except in the case of the default yellow emoji. While both groups associate this with a White identity, the effect was stronger in White participants. Our finding that emoji can index social variables will have experimental applications for researchers but also implications for designers: supposedly ``neutral`` defaults may be more representative of some users than others.

(2 min) This study investigates the performance of personalized automatic speech recognition (ASR) for recognizing disordered speech using small amounts of per-speaker adaptation data. We trained personalized models for 195 individuals with different types and severities of speech impairment with training sets ranging in size from <1 minute to 18-20 minutes of speech data. Word error rate (WER) thresholds were selected to determine Success Percentage (the percentage of personalized models reaching the target WER) in different application scenarios. For the home automation scenario, 79% of speakers reached the target WER with 18-20 minutes of speech; but even with only 3-4 minutes of speech, 63% of speakers reached the target WER. Further evaluation found similar improvement on test sets with conversational and out-of-domain, unprompted phrases. Our results demonstrate that with only a few minutes of recordings, individuals with disordered speech could benefit from personalized ASR.

(2 min) We propose a framework that learns to execute natural language instructions in an environment consisting of goal-reaching tasks that share components of their task descriptions. Our approach leverages the compositionality of both value functions and language, with the aim of reducing the sample complexity of learning novel tasks. First, we train a reinforcement learning agent to learn value functions that can be subsequently composed through a Boolean algebra to solve novel tasks. Second, we fine-tune a seq2seq model pretrained on web-scale corpora to map language to logical expressions that specify the required value function compositions. Evaluating our agent in the BabyAI domain, we observe a decrease of 86% in the number of training steps needed to learn a second task after mastering a single task. Results from ablation studies further indicate that it is the combination of compositional value functions and language representations that allows the agent to quickly generalize to new tasks.

(2 min) Transcripts generated by automatic speech recognition (ASR) systems for spoken documents lack structural annotations such as paragraphs, significantly reducing their readability. Automatically predicting paragraph segmentation for spoken documents may both improve readability and downstream NLP performance such as summarization and machine reading comprehension. We propose a sequence model with self-adaptive sliding window for accurate and efficient paragraph segmentation. We also propose an approach to exploit phonetic information, which significantly improves robustness of spoken document segmentation to ASR errors. Evaluations are conducted on the English Wiki-727K document segmentation benchmark, a Chinese Wikipedia-based document segmentation dataset we created, and an in-house Chinese spoken document dataset. Our proposed model outperforms the state-of-the-art (SOTA) model based on the same BERT-Base, increasing segmentation F1 on the English benchmark by 4.2 points and on Chinese datasets by 4.3-10.1 points, while reducing inference time to less than 1/6 of inference time of the current SOTA.

(2 min) Rumors are rampant in the era of social media. Conversation structures provide valuable clues to differentiate between real and fake claims. However, existing rumor detection methods are either limited to the strict relation of user responses or oversimplify the conversation structure. In this study, to substantially reinforces the interaction of user opinions while alleviating the negative impact imposed by irrelevant posts, we first represent the conversation thread as an undirected interaction graph. We then present a Claim-guided Hierarchical Graph Attention Network for rumor classification, which enhances the representation learning for responsive posts considering the entire social contexts and attends over the posts that can semantically infer the target claim. Extensive experiments on three Twitter datasets demonstrate that our rumor detection method achieves much better performance than state-of-the-art methods and exhibits a superior capacity for detecting rumors at early stages.

(2 min) Attention is an increasingly popular mechanism used in a wide range of neural architectures. The mechanism itself has been realized in a variety of formats. However, because of the fast-paced advances in this domain, a systematic overview of attention is still missing. In this article, we define a unified model for attention architectures in natural language processing, with a focus on those designed to work with vector representations of the textual data. We propose a taxonomy of attention models according to four dimensions: the representation of the input, the compatibility function, the distribution function, and the multiplicity of the input and/or output. We present the examples of how prior information can be exploited in attention models and discuss ongoing research efforts and open challenges in the area, providing the first extensive categorization of the vast body of literature in this exciting domain.

(2 min) Self-supervised pre-training has dramatically improved the performance of automatic speech recognition (ASR). However, most existing self-supervised pre-training approaches are task-agnostic, i.e., could be applied to various downstream tasks. And there is a gap between the task-agnostic pre-training and the task-specific downstream fine-tuning, which may degrade the downstream performance. In this work, we propose a novel pre-training paradigm called wav2vec-S, where we use task-specific semi-supervised pre-training to bridge this gap. Specifically, the semi-supervised pre-training is conducted on the basis of self-supervised pre-training such as wav2vec 2.0. Experiments on ASR show that compared to wav2vec 2.0, wav2vec-S only requires marginal increment of pre-training time but could significantly improve ASR performance on in-domain, cross-domain and cross-lingual datasets. The average relative WER reductions are 26.3% and 6.3% for 1h and 10h fine-tuning, respectively.

(2 min) NLP models that compare or consolidate information across multiple documents often struggle when challenged with recognizing substantial information redundancies across the texts. For example, in multi-document summarization it is crucial to identify salient information across texts and then generate a non-redundant summary, while facing repeated and usually differently-phrased salient content. To facilitate researching such challenges, the sentence-level task of \textit{sentence fusion} was proposed, yet previous datasets for this task were very limited in their size and scope. In this paper, we revisit and substantially extend previous dataset creation efforts. With careful modifications, relabeling and employing complementing data sources, we were able to triple the size of a notable earlier dataset. Moreover, we show that our extended version uses more representative texts for multi-document tasks and provides a larger and more diverse training set, which substantially improves model training.

(2 min) Sequence-to-sequence learning with neural networks has become the de facto standard for sequence prediction tasks. This approach typically models the local distribution over the next word with a powerful neural network that can condition on arbitrary context. While flexible and performant, these models often require large datasets for training and can fail spectacularly on benchmarks designed to test for compositional generalization. This work explores an alternative, hierarchical approach to sequence-to-sequence learning with quasi-synchronous grammars, where each node in the target tree is transduced by a node in the source tree. Both the source and target trees are treated as latent and induced during training. We develop a neural parameterization of the grammar which enables parameter sharing over the combinatorial space of derivation rules without the need for manual feature engineering. We apply this latent neural grammar to various domains -- a diagnostic language navigation task designed to test for compositional generalization (SCAN), style transfer, and small-scale machine translation -- and find that it performs respectably compared to standard baselines.

(2 min) Text discourse parsing weighs importantly in understanding information flow and argumentative structure in natural language, making it beneficial for downstream tasks. While previous work significantly improves the performance of RST discourse parsing, they are not readily applicable to practical use cases: (1) EDU segmentation is not integrated into most existing tree parsing frameworks, thus it is not straightforward to apply such models on newly-coming data. (2) Most parsers cannot be used in multilingual scenarios, because they are developed only in English. (3) Parsers trained from single-domain treebanks do not generalize well on out-of-domain inputs. In this work, we propose a document-level multilingual RST discourse parsing framework, which conducts EDU segmentation and discourse tree parsing jointly. Moreover, we propose a cross-translation augmentation strategy to enable the framework to support multilingual parsing and improve its domain generality. Experimental results show that our model achieves state-of-the-art performance on document-level multilingual RST parsing in all sub-tasks.

(2 min) Empathetic response generation aims to comprehend the user emotion and then respond to it appropriately. Most existing works merely focus on what the emotion is and ignore how the emotion is evoked, thus weakening the capacity of the model to understand the emotional experience of the user for generating empathetic responses. To tackle this problem, we consider the emotional causality, namely, what feelings the user expresses (i.e., emotion) and why the user has such feelings (i.e., cause). Then, we propose a novel graph-based model with multi-hop reasoning to model the emotional causality of the empathetic conversation. Finally, we demonstrate the effectiveness of our model on EMPATHETICDIALOGUES in comparison with several competitive models.

(2 min) Named Entity Recognition (NER) and Relation Extraction (RE) are the core sub-tasks for information extraction. Many recent works formulate these two tasks as the span (pair) classification problem, and thus focus on investigating how to obtain a better span representation from the pre-trained encoder. However, a major limitation of existing works is that they ignore the dependencies between spans (pairs). In this work, we propose a novel span representation approach, named Packed Levitated Markers, to consider the dependencies between the spans (pairs) by strategically packing the markers in the encoder. In particular, we propose a group packing strategy to enable our model to process massive spans together to consider their dependencies with limited resources. Furthermore, for those more complicated span pair classification tasks, we design a subject-oriented packing strategy, which packs each subject and all its objects into an instance to model the dependencies between the same-subject span pairs. Our experiments show that our model with packed levitated markers outperforms the sequence labeling model by 0.4%-1.9% F1 on three flat NER tasks, beats the token concat model on six NER benchmarks, and obtains a 3.5%-3.6% strict relation F1 improvement with higher speed over previous SOTA models on ACE04 and ACE05. Code and models are publicly available at https://github.com/thunlp/PL-Marker.

(2 min) Large-scale contrastive vision-language pre-training has shown significant progress in visual representation learning. Unlike traditional visual systems trained by a fixed set of discrete labels, a new paradigm was introduced in \cite{radford2021learning} to directly learn to align images with raw texts in an open-vocabulary setting. On downstream tasks, a carefully chosen text prompt is employed to make zero-shot predictions.~To avoid non-trivial prompt engineering, context optimization \cite{zhou2021coop} has been proposed to learn continuous vectors as task-specific prompts with few-shot training examples.~In this paper, we show that there is an alternative path to achieve better vision-language models other than prompt tuning.~While prompt tuning is for the textual inputs, we propose CLIP-Adapter to conduct fine-tuning with feature adapters on either visual or language branch. Specifically, CLIP-Adapter adopts an additional bottleneck layer to learn new features and performs residual-style feature blending with the original pre-trained features.~As a consequence, CLIP-Adapter is able to outperform context optimization while maintains a simple design. Experiments and extensive ablation studies on various visual classification tasks demonstrate the effectiveness of our approach.

(2 min) Why do models often attend to salient words, and how does this evolve throughout training? We approximate model training as a two stage process: early on in training when the attention weights are uniform, the model learns to translate individual input word `i` to `o` if they co-occur frequently. Later, the model learns to attend to `i` while the correct output is $o$ because it knows `i` translates to `o`. To formalize, we define a model property, Knowledge to Translate Individual Words (KTIW) (e.g. knowing that `i` translates to `o`), and claim that it drives the learning of the attention. This claim is supported by the fact that before the attention mechanism is learned, KTIW can be learned from word co-occurrence statistics, but not the other way around. Particularly, we can construct a training distribution that makes KTIW hard to learn, the learning of the attention fails, and the model cannot even learn the simple task of copying the input words to the output. Our approximation explains why models sometimes attend to salient words, and inspires a toy example where a multi-head attention model can overcome the above hard training distribution by improving learning dynamics rather than expressiveness. We end by discussing the limitation of our approximation framework and suggest future directions.

(2 min) We explore the link between the extent to which syntactic relations are preserved in translation and the ease of correctly constructing a parse tree in a zero-shot setting. While previous work suggests such a relation, it tends to focus on the macro level and not on the level of individual edges-a gap we aim to address. As a test case, we take the transfer of Universal Dependencies (UD) parsing from English to a diverse set of languages and conduct two sets of experiments. In one, we analyze zero-shot performance based on the extent to which English source edges are preserved in translation. In another, we apply three linguistically motivated transformations to UD, creating more cross-lingually stable versions of it, and assess their zero-shot parsability. In order to compare parsing performance across different schemes, we perform extrinsic evaluation on the downstream task of cross-lingual relation extraction (RE) using a subset of a popular English RE benchmark translated to Russian and Korean. In both sets of experiments, our results suggest a strong relation between cross-lingual stability and zero-shot parsing performance.

(2 min) As deep networks begin to be deployed as autonomous agents, the issue of how they can communicate with each other becomes important. Here, we train two deep nets from scratch to perform realistic referent identification through unsupervised emergent communication. We show that the largely interpretable emergent protocol allows the nets to successfully communicate even about object types they did not see at training time. The visual representations induced as a by-product of our training regime, moreover, show comparable quality, when re-used as generic visual features, to a recent self-supervised learning model. Our results provide concrete evidence of the viability of (interpretable) emergent deep net communication in a more realistic scenario than previously considered, as well as establishing an intriguing link between this field and self-supervised visual learning.

(2 min) The ability to continuously expand knowledge over time and utilize it to rapidly generalize to new tasks is a key feature of human linguistic intelligence. Existing models that pursue rapid generalization to new tasks (e.g., few-shot learning methods), however, are mostly trained in a single shot on fixed datasets, unable to dynamically expand their knowledge; while continual learning algorithms are not specifically designed for rapid generalization. We present a new learning setup, Continual Learning of Few-Shot Learners (CLIF), to address the challenges of both learning settings in a unified setup. CLIF assumes a model learns from a sequence of diverse NLP tasks arriving sequentially, accumulating knowledge for improved generalization to new tasks, while also retaining performance on the tasks learned earlier. We examine how the generalization ability is affected in the continual learning setup, evaluate a number of continual learning algorithms, and propose a novel regularized adapter generation approach. We find that catastrophic forgetting affects generalization ability to a less degree than performance on seen tasks; while continual learning algorithms can still bring considerable benefit to the generalization ability.

(2 min) Self-supervised pretraining on speech data has achieved a lot of progress. High-fidelity representation of the speech signal is learned from a lot of untranscribed data and shows promising performance. Recently, there are several works focusing on evaluating the quality of self-supervised pretrained representations on various tasks without domain restriction, e.g. SUPERB. However, such evaluations do not provide a comprehensive comparison among many ASR benchmark corpora. In this paper, we focus on the general applications of pretrained speech representations, on advanced end-to-end automatic speech recognition (E2E-ASR) models. We select several pretrained speech representations and present the experimental results on various open-source and publicly available corpora for E2E-ASR. Without any modification of the back-end model architectures or training strategy, some of the experiments with pretrained representations, e.g., WSJ, WSJ0-2mix with HuBERT, reach or outperform current state-of-the-art (SOTA) recognition performance. Moreover, we further explore more scenarios for whether the pretraining representations are effective, such as the cross-language or overlapped speech. The scripts, configuratons and the trained models have been released in ESPnet to let the community reproduce our experiments and improve them.

(2 min) Machine translation has wide applications in daily life. In mission-critical applications such as translating official documents, incorrect translation can have unpleasant or sometimes catastrophic consequences. This motivates recent research on testing methodologies for machine translation systems. Existing methodologies mostly rely on metamorphic relations designed at the textual level (e.g., Levenshtein distance) or syntactic level (e.g., the distance between grammar structures) to determine the correctness of translation results. However, these metamorphic relations do not consider whether the original and translated sentences have the same meaning (i.e., Semantic similarity). Therefore, in this paper, we propose SemMT, an automatic testing approach for machine translation systems based on semantic similarity checking. SemMT applies round-trip translation and measures the semantic similarity between the original and translated sentences. Our insight is that the semantics expressed by the logic and numeric constraint in sentences can be captured using regular expressions (or deterministic finite automata) where efficient equivalence/similarity checking algorithms are available. Leveraging the insight, we propose three semantic similarity metrics and implement them in SemMT. The experiment result reveals SemMT can achieve higher effectiveness compared with state-of-the-art works, achieving an increase of 21% and 23% on accuracy and F-Score, respectively. We also explore potential improvements that can be achieved when proper combinations of metrics are adopted. Finally, we discuss a solution to locate the suspicious trip in round-trip translation, which may shed lights on further exploration.

(2 min) As neural-network-based QA models become deeper and more complex, there is a demand for robust frameworks which can access a model's rationale for its prediction. Current techniques that provide insights on a model's working are either dependent on adversarial datasets or are proposing models with explicit explanation generation components. These techniques are time-consuming and challenging to extend to existing models and new datasets. In this work, we use `Integrated Gradients' to extract rationale for existing state-of-the-art models in the task of Reading Comprehension based Question Answering (RCQA). On detailed analysis and comparison with collected human rationales, we find that though ~40-80% words of extracted rationale coincide with the human rationale (precision), only 6-19% of human rationale is present in the extracted rationale (recall).

(2 min) Automated audio captioning aims at generating textual descriptions for an audio clip. To evaluate the quality of generated audio captions, previous works directly adopt image captioning metrics like SPICE and CIDEr, without justifying their suitability in this new domain, which may mislead the development of advanced models. This problem is still unstudied due to the lack of human judgment datasets on caption quality. Therefore, we firstly construct two evaluation benchmarks, AudioCaps-Eval and Clotho-Eval. They are established with pairwise comparison instead of absolute rating to achieve better inter-annotator agreement. Current metrics are found in poor correlation with human annotations on these datasets. To overcome their limitations, we propose a metric named FENSE, where we combine the strength of Sentence-BERT in capturing similarity, and a novel Error Detector to penalize erroneous sentences for robustness. On the newly established benchmarks, FENSE outperforms current metrics by 14-25% accuracy. Code, data and web demo available at: https://github.com/blmoistawinde/fense

(2 min) Recent progress on end-to-end neural diarization (EEND) has enabled overlap-aware speaker diarization with a single neural network. This paper proposes to enhance EEND by using multi-channel signals from distributed microphones. We replace Transformer encoders in EEND with two types of encoders that process a multi-channel input: spatio-temporal and co-attention encoders. Both are independent of the number and geometry of microphones and suitable for distributed microphone settings. We also propose a model adaptation method using only single-channel recordings. With simulated and real-recorded datasets, we demonstrated that the proposed method outperformed conventional EEND when a multi-channel input was given while maintaining comparable performance with a single-channel input. We also showed that the proposed method performed well even when spatial information is inoperative given multi-channel inputs, such as in hybrid meetings in which the utterances of multiple remote participants are played back from the same loudspeaker.

(2 min) We introduce dense relational captioning, a novel image captioning task which aims to generate multiple captions with respect to relational information between objects in a visual scene. Relational captioning provides explicit descriptions for each relationship between object combinations. This framework is advantageous in both diversity and amount of information, leading to a comprehensive image understanding based on relationships, e.g., relational proposal generation. For relational understanding between objects, the part-of-speech (POS; i.e., subject-object-predicate categories) can be a valuable prior information to guide the causal sequence of words in a caption. We enforce our framework to learn not only to generate captions but also to understand the POS of each word. To this end, we propose the multi-task triple-stream network (MTTSNet) which consists of three recurrent units responsible for each POS which is trained by jointly predicting the correct captions and POS for each word. In addition, we found that the performance of MTTSNet can be improved by modulating the object embeddings with an explicit relational module. We demonstrate that our proposed model can generate more diverse and richer captions, via extensive experimental analysis on large scale datasets and several metrics. Then, we present applications of our framework to holistic image captioning, scene graph generation, and retrieval tasks.

(2 min) Several studies have explored various advantages of multilingual pre-trained models (e.g., multilingual BERT) in capturing shared linguistic knowledge. However, their limitations have not been paid enough attention. In this paper, we investigate the representation degeneration problem in multilingual contextual word representations (CWRs) of BERT and show that the embedding spaces of the selected languages suffer from anisotropy problem. Our experimental results demonstrate that, similarly to their monolingual counterparts, increasing the isotropy of multilingual embedding space can significantly improve its representation power and performance. Our analysis indicates that although the degenerated directions vary in different languages, they encode similar linguistic knowledge, suggesting a shared linguistic space among languages.

(2 min) Natural language descriptions sometimes accompany visualizations to better communicate and contextualize their insights, and to improve their accessibility for readers with disabilities. However, it is difficult to evaluate the usefulness of these descriptions, and how effectively they improve access to meaningful information, because we have little understanding of the semantic content they convey, and how different readers receive this content. In response, we introduce a conceptual model for the semantic content conveyed by natural language descriptions of visualizations. Developed through a grounded theory analysis of 2,147 sentences, our model spans four levels of semantic content: enumerating visualization construction properties (e.g., marks and encodings); reporting statistical concepts and relations (e.g., extrema and correlations); identifying perceptual and cognitive phenomena (e.g., complex trends and patterns); and elucidating domain-specific insights (e.g., social and political context). To demonstrate how our model can be applied to evaluate the effectiveness of visualization descriptions, we conduct a mixed-methods evaluation with 30 blind and 90 sighted readers, and find that these reader groups differ significantly on which semantic content they rank as most useful. Together, our model and findings suggest that access to meaningful information is strongly reader-specific, and that research in automatic visualization captioning should orient toward descriptions that more richly communicate overall trends and statistics, sensitive to reader preferences. Our work further opens a space of research on natural language as a data interface coequal with visualization.

(2 min) In this paper, we introduce the Eval4NLP-2021shared task on explainable quality estimation. Given a source-translation pair, this shared task requires not only to provide a sentence-level score indicating the overall quality of the translation, but also to explain this score by identifying the words that negatively impact translation quality. We present the data, annotation guidelines and evaluation setup of the shared task, describe the six participating systems, and analyze the results. To the best of our knowledge, this is the first shared task on explainable NLP evaluation metrics. Datasets and results are available at https://github.com/eval4nlp/SharedTask2021.

(2 min) The rise of Internet has made it a major source of information. Unfortunately, not all information online is true, and thus a number of fact-checking initiatives have been launched, both manual and automatic, to deal with the problem. Here, we present our contribution in this regard: \emph{WhatTheWikiFact}, a system for automatic claim verification using Wikipedia. The system can predict the veracity of an input claim, and it further shows the evidence it has retrieved as part of the verification process. It shows confidence scores and a list of relevant Wikipedia articles, together with detailed information about each article, including the phrase used to retrieve it, the most relevant sentences extracted from it and their stance with respect to the input claim, as well as the associated probabilities. The system supports several languages: Bulgarian, English, and Russian.

(2 min) There is mounting evidence that existing neural network models, in particular the very popular sequence-to-sequence architecture, struggle with compositional generalization, i.e., the ability to systematically generalize to unseen compositions of seen components. In this paper we demonstrate that one of the reasons hindering compositional generalization relates to the representations being entangled. We propose an extension to sequence-to-sequence models which allows us to learn disentangled representations by adaptively re-encoding (at each time step) the source input. Specifically, we condition the source representations on the newly decoded target context which makes it easier for the encoder to exploit specialized information for each prediction rather than capturing all source information in a single forward pass. Experimental results on semantic parsing and machine translation empirically show that our proposal yields more disentangled representations and better generalization.

(2 min) Evaluation metrics are a key ingredient for progress of text generation systems. In recent years, several BERT-based evaluation metrics have been proposed (including BERTScore, MoverScore, BLEURT, etc.) which correlate much better with human assessment of text generation quality than BLEU or ROUGE, invented two decades ago. However, little is known what these metrics, which are based on black-box language model representations, actually capture (it is typically assumed they model semantic similarity). In this work, we \wei{use a simple regression based global explainability technique to} disentangle metric scores along linguistic factors, including semantics, syntax, morphology, and lexical overlap. We show that the different metrics capture all aspects to some degree, but that they are all substantially sensitive to lexical overlap, just like BLEU and ROUGE. This exposes limitations of these novelly proposed metrics, which we also highlight in an adversarial test scenario.

(2 min) Cognitively inspired Natural Language Pro-cessing uses human-derived behavioral datalike eye-tracking data, which reflect the seman-tic representations of language in the humanbrain to augment the neural nets to solve arange of tasks spanning syntax and semanticswith the aim of teaching machines about lan-guage processing mechanisms. In this paper,we use the ZuCo 1.0 and ZuCo 2.0 dataset con-taining the eye-gaze features to explore differ-ent linguistic models to directly predict thesegaze features for each word with respect to itssentence. We tried different neural networkmodels with the words as inputs to predict thetargets. And after lots of experimentation andfeature engineering finally devised a novel ar-chitecture consisting of RoBERTa Token Clas-sifier with a dense layer on top for languagemodeling and a stand-alone model consistingof dense layers followed by a transformer layerfor the extra features we engineered. Finally,we took the mean of the outputs of both thesemodels to make the final predictions. We eval-uated the models using mean absolute error(MAE) and the R2 score for each target.

(2 min) Pretraining Neural Language Models (NLMs) over a large corpus involves chunking the text into training examples, which are contiguous text segments of sizes processable by the neural architecture. We highlight a bias introduced by this common practice: we prove that the pretrained NLM can model much stronger dependencies between text segments that appeared in the same training example, than it can between text segments that appeared in different training examples. This intuitive result has a twofold role. First, it formalizes the motivation behind a broad line of recent successful NLM training heuristics, proposed for the pretraining and fine-tuning stages, which do not necessarily appear related at first glance. Second, our result clearly indicates further improvements to be made in NLM pretraining for the benefit of Natural Language Understanding tasks. As an example, we propose "kNN-Pretraining": we show that including semantically related non-neighboring sentences in the same pretraining example yields improved sentence representations and open domain question answering abilities. This theoretically motivated degree of freedom for "pretraining example design" indicates new training schemes for self-improving representations.

(2 min) Distantly supervised named entity recognition (DS-NER) efficiently reduces labor costs but meanwhile intrinsically suffers from the label noise due to the strong assumption of distant supervision. Typically, the wrongly labeled instances comprise numbers of incomplete and inaccurate annotation noise, while most prior denoising works are only concerned with one kind of noise and fail to fully explore useful information in the whole training set. To address this issue, we propose a robust learning paradigm named Self-Collaborative Denoising Learning (SCDL), which jointly trains two teacher-student networks in a mutually-beneficial manner to iteratively perform noisy label refinery. Each network is designed to exploit reliable labels via self denoising, and two networks communicate with each other to explore unreliable annotations by collaborative denoising. Extensive experimental results on five real-world datasets demonstrate that SCDL is superior to state-of-the-art DS-NER denoising methods.

(2 min) Sequence expansion between encoder and decoder is a critical challenge in sequence-to-sequence TTS. Attention-based methods achieve great naturalness but suffer from unstable issues like missing and repeating phonemes, not to mention accurate duration control. Duration-informed methods, on the contrary, seem to easily adjust phoneme duration but show obvious degradation in speech naturalness. This paper proposes PAMA-TTS to address the problem. It takes the advantage of both flexible attention and explicit duration models. Based on the monotonic attention mechanism, PAMA-TTS also leverages token duration and relative position of a frame, especially countdown information, i.e. in how many future frames the present phoneme will end. They help the attention to move forward along the token sequence in a soft but reliable control. Experimental results prove that PAMA-TTS achieves the highest naturalness, while has on-par or even better duration controllability than the duration-informed model.

(2 min) Position encoding in transformer architecture provides supervision for dependency modeling between elements at different positions in the sequence. We investigate various methods to encode positional information in transformer-based language models and propose a novel implementation named Rotary Position Embedding(RoPE). The proposed RoPE encodes absolute positional information with rotation matrix and naturally incorporates explicit relative position dependency in self-attention formulation. Notably, RoPE comes with valuable properties such as flexibility of being expand to any sequence lengths, decaying inter-token dependency with increasing relative distances, and capability of equipping the linear self-attention with relative position encoding. As a result, the enhanced transformer with rotary position embedding, or RoFormer, achieves superior performance in tasks with long texts. We release the theoretical analysis along with some preliminary experiment results on Chinese data. The undergoing experiment for English benchmark will soon be updated.

(2 min) Current Open-Domain Question Answering (ODQA) model paradigm often contains a retrieving module and a reading module. Given an input question, the reading module predicts the answer from the relevant passages which are retrieved by the retriever. The recent proposed Fusion-in-Decoder (FiD), which is built on top of the pretrained generative model T5, achieves the state-of-the-art performance in the reading module. Although being effective, it remains constrained by inefficient attention on all retrieved passages which contain a lot of noise. In this work, we propose a novel method KG-FiD, which filters noisy passages by leveraging the structural relationship among the retrieved passages with a knowledge graph. We initiate the passage node embedding from the FiD encoder and then use graph neural network (GNN) to update the representation for reranking. To improve the efficiency, we build the GNN on top of the intermediate layer output of the FiD encoder and only pass a few top reranked passages into the higher layers of encoder and decoder for answer generation. We also apply the proposed GNN based reranking method to enhance the passage retrieval results in the retrieving module. Extensive experiments on common ODQA benchmark datasets (Natural Question and TriviaQA) demonstrate that KG-FiD can improve vanilla FiD by up to 1.5% on answer exact match score and achieve comparable performance with FiD with only 40% of computation cost.

(2 min) Summarization systems are ultimately evaluated by human annotators and raters. Usually, annotators and raters do not reflect the demographics of end users, but are recruited through student populations or crowdsourcing platforms with skewed demographics. For two different evaluation scenarios -- evaluation against gold summaries and system output ratings -- we show that summary evaluation is sensitive to protected attributes. This can severely bias system development and evaluation, leading us to build models that cater for some groups rather than others.

(2 min) We investigate the dynamics of increasing the number of model parameters versus the number of labeled examples across a wide variety of tasks. Our exploration reveals that while scaling parameters consistently yields performance improvements, the contribution of additional examples highly depends on the task's format. Specifically, in open question answering tasks, enlarging the training set does not improve performance. In contrast, classification, extractive question answering, and multiple choice tasks benefit so much from additional examples that collecting a few hundred examples is often "worth" billions of parameters. We hypothesize that unlike open question answering, which involves recalling specific information, solving strategies for tasks with a more restricted output space transfer across examples, and can therefore be learned with small amounts of labeled data.

(2 min) Fine-tuning large pre-trained language models on downstream tasks has become the de-facto learning paradigm in NLP. However, conventional approaches fine-tune all the parameters of the pre-trained model, which becomes prohibitive as the model size and the number of tasks grow. Recent work has proposed a variety of parameter-efficient transfer learning methods that only fine-tune a small number of (extra) parameters to attain strong performance. While effective, the critical ingredients for success and the connections among the various methods are poorly understood. In this paper, we break down the design of state-of-the-art parameter-efficient transfer learning methods and present a unified framework that establishes connections between them. Specifically, we re-frame them as modifications to specific hidden states in pre-trained models, and define a set of design dimensions along which different methods vary, such as the function to compute the modification and the position to apply the modification. Through comprehensive empirical studies across machine translation, text summarization, language understanding, and text classification benchmarks, we utilize the unified view to identify important design choices in previous methods. Furthermore, our unified framework enables the transfer of design elements across different approaches, and as a result we are able to instantiate new parameter-efficient fine-tuning methods that tune less parameters than previous methods while being more effective, achieving comparable results to fine-tuning all parameters on all four tasks.

(2 min) We have a FPGA design, we make it fast, efficient, and tested for a few important examples. Now we must infer a general solution to deploy in the data center. Here, we describe the FPGA DPUV3INT8 design and our compiler effort. The hand-tuned SW-HW solution for Resnet50\_v1 has (close to) 2 times better images per second (throughput) than our best FPGA implementation; the compiler generalizes the hand written techniques achieving about 1.5 times better performance for the same example, the compiler generalizes the optimizations to a model zoo of networks, and it achieves 80+\% HW efficiency.

(2 min) Deep neural networks are prone to adversarial examples that maliciously alter the network's outcome. Due to the increasing popularity of 3D sensors in safety-critical systems and the vast deployment of deep learning models for 3D point sets, there is a growing interest in adversarial attacks and defenses for such models. So far, the research has focused on the semantic level, namely, deep point cloud classifiers. However, point clouds are also widely used in a geometric-related form that includes encoding and reconstructing the geometry. In this work, we are the first to consider the problem of adversarial examples at a geometric level. In this setting, the question is how to craft a small change to a clean source point cloud that leads, after passing through an autoencoder model, to the reconstruction of a different target shape. Our attack is in sharp contrast to existing semantic attacks on 3D point clouds. While such works aim to modify the predicted label by a classifier, we alter the entire reconstructed geometry. Additionally, we demonstrate the robustness of our attack in the case of defense, where we show that remnant characteristics of the target shape are still present at the output after applying the defense to the adversarial input. Our code is publicly available at https://github.com/itailang/geometric_adv.

(2 min) We propose a fully differentiable architecture for simultaneous semantic and instance segmentation (a.k.a. panoptic segmentation) consisting of a convolutional neural network and an asymmetric multiway cut problem solver. The latter solves a combinatorial optimization problem that elegantly incorporates semantic and boundary predictions to produce a panoptic labeling. Our formulation allows to directly maximize a smooth surrogate of the panoptic quality metric by backpropagating the gradient through the optimization problem. Experimental evaluation shows improvement by backpropagating through the optimization problem w.r.t. comparable approaches on Cityscapes and COCO datasets. Overall, our approach shows the utility of using combinatorial optimization in tandem with deep learning in a challenging large scale real-world problem and showcases benefits and insights into training such an architecture.

(2 min) In this paper, we present a neat yet effective transformer-based framework for visual grounding, namely TransVG, to address the task of grounding a language query to the corresponding region onto an image. The state-of-the-art methods, including two-stage or one-stage ones, rely on a complex module with manually-designed mechanisms to perform the query reasoning and multi-modal fusion. However, the involvement of certain mechanisms in fusion module design, such as query decomposition and image scene graph, makes the models easily overfit to datasets with specific scenarios, and limits the plenitudinous interaction between the visual-linguistic context. To avoid this caveat, we propose to establish the multi-modal correspondence by leveraging transformers, and empirically show that the complex fusion modules (\eg, modular attention network, dynamic graph, and multi-modal tree) can be replaced by a simple stack of transformer encoder layers with higher performance. Moreover, we re-formulate the visual grounding as a direct coordinates regression problem and avoid making predictions out of a set of candidates (\emph{i.e.}, region proposals or anchor boxes). Extensive experiments are conducted on five widely used datasets, and a series of state-of-the-art records are set by our TransVG. We build the benchmark of transformer-based visual grounding framework and make the code available at \url{https://github.com/djiajunustc/TransVG}.

(2 min) Convolutional neural networks (CNNs) have made significant advancement, however, they are widely known to be vulnerable to adversarial attacks. Adversarial training is the most widely used technique for improving adversarial robustness to strong white-box attacks. Prior works have been evaluating and improving the model average robustness without class-wise evaluation. The average evaluation alone might provide a false sense of robustness. For example, the attacker can focus on attacking the vulnerable class, which can be dangerous, especially, when the vulnerable class is a critical one, such as "human" in autonomous driving. We propose an empirical study on the class-wise accuracy and robustness of adversarially trained models. We find that there exists inter-class discrepancy for accuracy and robustness even when the training dataset has an equal number of samples for each class. For example, in CIFAR10, "cat" is much more vulnerable than other classes. Moreover, this inter-class discrepancy also exists for normally trained models, while adversarial training tends to further increase the discrepancy. Our work aims to investigate the following questions: (a) is the phenomenon of inter-class discrepancy universal regardless of datasets, model architectures and optimization hyper-parameters? (b) If so, what can be possible explanations for the inter-class discrepancy? (c) Can the techniques proposed in the long tail classification be readily extended to adversarial training for addressing the inter-class discrepancy?

(2 min) The recognition of unseen objects from a semantic representation or textual description, usually denoted as zero-shot learning, is more prone to be used in real-world scenarios when compared to traditional object recognition. Nevertheless, no work has evaluated the feasibility of deploying zero-shot learning approaches in these scenarios, particularly when using low-power devices. In this paper, we provide the first benchmark on the inference time of zero-shot learning, comprising an evaluation of state-of-the-art approaches regarding their speed/accuracy trade-off. An analysis to the processing time of the different phases of the ZSL inference stage reveals that visual feature extraction is the major bottleneck in this paradigm, but, we show that lightweight networks can dramatically reduce the overall inference time without reducing the accuracy obtained by the de facto ResNet101 architecture. Also, this benchmark evaluates how different ZSL approaches perform in low-power devices, and how the visual feature extraction phase could be optimized in this hardware. To foster the research and deployment of ZSL systems capable of operating in real-world scenarios, we release the evaluation framework used in this benchmark (https://github.com/CristianoPatricio/zsl-methods).

(0 min) A key reason for the lack of reliability of deep neural networks in the real world is their heavy reliance on {\it spurious} input features that are causally unrelated to the true label. Focusing on image classifications, we define causal attributes as the set of visual features that are always a part of the object while spurious attributes are the ones that are likely to {\it co-occur} with the object but not a part of it (e.g., attribute ``fingers" for class ``band aid"). Traditional methods for discovering spurious features either require extensive human annotations (thus, not scalable), or are useful on specific models. In this work, we introduce a {\it scalable} framework to discover a subset of spurious and causal visual attributes used in inferences of a general model and localize them on a large number of images with minimal human supervision. Our methodology is based on this key idea: to identify spurious or causal \textit{visual attributes} used in model predictions, we identify spurious or causal \textit{neural features} (penultimate layer neurons of a robust model) via limited human supervision (e.g., using top 5 activating images per feature). We then show that these neural feature annotations {\it generalize} extremely well to many more images {\it without} any human supervision. We use the activation maps for these neural features as the soft masks to highlight spurious or causal visual attributes. Using this methodology, we introduce the {\it Causal Imagenet} dataset containing causal and spurious masks for a large set of samples from Imagenet. We assess the performance of several popular Imagenet models and show that they rely heavily on various spurious features in their predictions.

(0 min) Rigid registration of partial observations is a fundamental problem in various applied fields. In computer graphics, special attention has been given to the registration between two partial point clouds generated by scanning devices. State-of-the-art registration techniques still struggle when the overlap region between the two point clouds is small, and completely fail if there is no overlap between the scan pairs. In this paper, we present a learning-based technique that alleviates this problem, and allows registration between point clouds, presented in arbitrary poses, and having little or even no overlap, a setting that has been referred to as tele-registration. Our technique is based on a novel neural network design that learns a prior of a class of shapes and can complete a partial shape. The key idea is combining the registration and completion tasks in a way that reinforces each other. In particular, we simultaneously train the registration network and completion network using two coupled flows, one that register-and-complete, and one that complete-and-register, and encourage the two flows to produce a consistent result. We show that, compared with each separate flow, this two-flow training leads to robust and reliable tele-registration, and hence to a better point cloud prediction that completes the registered scans. It is also worth mentioning that each of the components in our neural network outperforms state-of-the-art methods in both completion and registration. We further analyze our network with several ablation studies and demonstrate its performance on a large number of partial point clouds, both synthetic and real-world, that have only small or no overlap.

(3 min) Boundary Representations (B-Reps) are the industry standard in 3D Computer Aided Design/Manufacturing (CAD/CAM) and industrial design due to their fidelity in representing stylistic details. However, they have been ignored in the 3D style research. Existing 3D style metrics typically operate on meshes or pointclouds, and fail to account for end-user subjectivity by adopting fixed definitions of style, either through crowd-sourcing for style labels or hand-crafted features. We propose UVStyle-Net, a style similarity measure for B-Reps that leverages the style signals in the second order statistics of the activations in a pre-trained (unsupervised) 3D encoder, and learns their relative importance to a subjective end-user through few-shot learning. Our approach differs from all existing data-driven 3D style methods since it may be used in completely unsupervised settings, which is desirable given the lack of publicly available labelled B-Rep datasets. More importantly, the few-shot learning accounts for the inherent subjectivity associated with style. We show quantitatively that our proposed method with B-Reps is able to capture stronger style signals than alternative methods on meshes and pointclouds despite its significantly greater computational efficiency. We also show it is able to generate meaningful style gradients with respect to the input shape, and that few-shot learning with as few as two positive examples selected by an end-user is sufficient to significantly improve the style measure. Finally, we demonstrate its efficacy on a large unlabeled public dataset of CAD models. Source code and data will be released in the future.

(2 min) Robust face reconstruction from monocular image in general lighting conditions is challenging. Methods combining deep neural network encoders with differentiable rendering have opened up the path for very fast monocular reconstruction of geometry, lighting and reflectance. They can also be trained in self-supervised manner for increased robustness and better generalization. However, their differentiable rasterization based image formation models, as well as underlying scene parameterization, limit them to Lambertian face reflectance and to poor shape details. More recently, ray tracing was introduced for monocular face reconstruction within a classic optimization-based framework and enables state-of-the art results. However optimization-based approaches are inherently slow and lack robustness. In this paper, we build our work on the aforementioned approaches and propose a new method that greatly improves reconstruction quality and robustness in general scenes. We achieve this by combining a CNN encoder with a differentiable ray tracer, which enables us to base the reconstruction on much more advanced personalized diffuse and specular albedos, a more sophisticated illumination model and a plausible representation of self-shadows. This enables to take a big leap forward in reconstruction quality of shape, appearance and lighting even in scenes with difficult illumination. With consistent face attributes reconstruction, our method leads to practical applications such as relighting and self-shadows removal. Compared to state-of-the-art methods, our results show improved accuracy and validity of the approach.

(2 min) Neural implicit surface representations have emerged as a promising paradigm to capture 3D shapes in a continuous and resolution-independent manner. However, adapting them to articulated shapes is non-trivial. Existing approaches learn a backward warp field that maps deformed to canonical points. However, this is problematic since the backward warp field is pose dependent and thus requires large amounts of data to learn. To address this, we introduce SNARF, which combines the advantages of linear blend skinning (LBS) for polygonal meshes with those of neural implicit surfaces by learning a forward deformation field without direct supervision. This deformation field is defined in canonical, pose-independent space, allowing for generalization to unseen poses. Learning the deformation field from posed meshes alone is challenging since the correspondences of deformed points are defined implicitly and may not be unique under changes of topology. We propose a forward skinning model that finds all canonical correspondences of any deformed point using iterative root finding. We derive analytical gradients via implicit differentiation, enabling end-to-end training from 3D meshes with bone transformations. Compared to state-of-the-art neural implicit representations, our approach generalizes better to unseen poses while preserving accuracy. We demonstrate our method in challenging scenarios on (clothed) 3D humans in diverse and unseen poses.

(2 min) We explore the zero-shot setting for day-night domain adaptation. The traditional domain adaptation setting is to train on one domain and adapt to the target domain by exploiting unlabeled data samples from the test set. As gathering relevant test data is expensive and sometimes even impossible, we remove any reliance on test data imagery and instead exploit a visual inductive prior derived from physics-based reflection models for domain adaptation. We cast a number of color invariant edge detectors as trainable layers in a convolutional neural network and evaluate their robustness to illumination changes. We show that the color invariant layer reduces the day-night distribution shift in feature map activations throughout the network. We demonstrate improved performance for zero-shot day to night domain adaptation on both synthetic as well as natural datasets in various tasks, including classification, segmentation and place recognition.

(2 min) Robot-assisted surgery is an established clinical practice. The automatic identification of surgical actions is needed for a range of applications, including performance assessment of trainees and surgical process modeling for autonomous execution and monitoring. However, supervised action identification is not feasible, due to the burden of manually annotating recordings of potentially complex and long surgical executions. Moreover, often few example executions of a surgical procedure can be recorded. This paper proposes a novel fast algorithm for unsupervised identification of surgical actions in a standard surgical training task, the ring transfer, executed with da Vinci Research Kit. Exploiting kinematic and semantic visual features automatically extracted from a very limited dataset of executions, we are able to significantly outperform state-of-the-art results on a dataset of non-expert executions (58\% vs. 24\% F1-score), and improve performance in the presence of noise, short actions and non-homogeneous workflows, i.e. non repetitive action sequences.

(2 min) Recent works on learned image compression perform encoding and decoding processes in a full-resolution manner, resulting in two problems when deployed for practical applications. First, parallel acceleration of the autoregressive entropy model cannot be achieved due to serial decoding. Second, full-resolution inference often causes the out-of-memory(OOM) problem with limited GPU resources, especially for high-resolution images. Block partition is a good design choice to handle the above issues, but it brings about new challenges in reducing the redundancy between blocks and eliminating block effects. To tackle the above challenges, this paper provides a learned block-based hybrid image compression (LBHIC) framework. Specifically, we introduce explicit intra prediction into a learned image compression framework to utilize the relation among adjacent blocks. Superior to context modeling by linear weighting of neighbor pixels in traditional codecs, we propose a contextual prediction module (CPM) to better capture long-range correlations by utilizing the strip pooling to extract the most relevant information in neighboring latent space, thus achieving effective information prediction. Moreover, to alleviate blocking artifacts, we further propose a boundary-aware postprocessing module (BPM) with the edge importance taken into account. Extensive experiments demonstrate that the proposed LBHIC codec outperforms the VVC, with a bit-rate conservation of 4.1%, and reduces the decoding time by approximately 86.7% compared with that of state-of-the-art learned image compression methods.

(0 min) Articulated objects exist widely in the real world. However, previous 3D generative methods for unsupervised part decomposition are unsuitable for such objects, because they assume a spatially fixed part location, resulting in inconsistent part parsing. In this paper, we propose PPD (unsupervised Pose-aware Part Decomposition) to address a novel setting that explicitly targets man-made articulated objects with mechanical joints, considering the part poses. We show that category-common prior learning for both part shapes and poses facilitates the unsupervised learning of (1) part decomposition with non-primitive-based implicit representation, and (2) part pose as joint parameters under single-frame shape supervision. We evaluate our method on synthetic and real datasets, and we show that it outperforms previous works in consistent part parsing of the articulated objects based on comparable part pose estimation performance to the supervised baseline.

(0 min) Model-agnostic meta-learning (MAML) is arguably one of the most popular meta-learning algorithms nowadays. Nevertheless, its performance on few-shot classification is far behind many recent algorithms dedicated to the problem. In this paper, we point out several key facets of how to train MAML to excel in few-shot classification. First, we find that MAML needs a large number of gradient steps in its inner loop update, which contradicts its common usage in few-shot classification. Second, we find that MAML is sensitive to the class label assignments during meta-testing. Concretely, MAML meta-trains the initialization of an $N$-way classifier. These $N$ ways, during meta-testing, then have $N!$ different permutations to be paired with a few-shot task of $N$ novel classes. We find that these permutations lead to a huge variance of accuracy, making MAML unstable in few-shot classification. Third, we investigate several approaches to make MAML permutation-invariant, among which meta-training a single vector to initialize all the $N$ weight vectors in the classification head performs the best. On benchmark datasets like MiniImageNet and TieredImageNet, our approach, which we name UNICORN-MAML, performs on a par with or even outperforms state-of-the-art few-shot classification algorithms, without sacrificing MAML's simplicity.

(0 min) We introduce a novel and uniform framework for quantum pixel representations that overarches many of the most popular representations proposed in the recent literature, such as (I)FRQI, (I)NEQR, MCRQI, and (I)NCQI. The proposed QPIXL framework results in more efficient circuit implementations and significantly reduces the gate complexity for all considered quantum pixel representations. Our method only requires a linear number of gates in terms of the number of pixels and does not use ancilla qubits. Furthermore, the circuits only consist of Ry gates and CNOT gates making them practical in the NISQ era. Additionally, we propose a circuit and image compression algorithm that is shown to be highly effective, being able to reduce the necessary gates to prepare an FRQI state for example scientific images by up to 90% without sacrificing image quality. Our algorithms are made publicly available as part of QPIXL++, a Quantum Image Pixel Library.

(0 min) This article aims to study the topological invariant properties encoded in node graph representational embeddings by utilizing tools available in persistent homology. Specifically, given a node embedding representation algorithm, we consider the case when these embeddings are real-valued. By viewing these embeddings as scalar functions on a domain of interest, we can utilize the tools available in persistent homology to study the topological information encoded in these representations. Our construction effectively defines a unique persistence-based graph descriptor, on both the graph and node levels, for every node representation algorithm. To demonstrate the effectiveness of the proposed method, we study the topological descriptors induced by DeepWalk, Node2Vec and Diff2Vec.

(0 min) Many existing approaches for unsupervised domain adaptation (UDA) focus on adapting under only data distribution shift and offer limited success under additional cross-domain label distribution shift. Recent work based on self-training using target pseudo-labels has shown promise, but on challenging shifts pseudo-labels may be highly unreliable, and using them for self-training may cause error accumulation and domain misalignment. We propose Selective Entropy Optimization via Committee Consistency (SENTRY), a UDA algorithm that judges the reliability of a target instance based on its predictive consistency under a committee of random image transformations. Our algorithm then selectively minimizes predictive entropy to increase confidence on highly consistent target instances, while maximizing predictive entropy to reduce confidence on highly inconsistent ones. In combination with pseudo-label based approximate target class balancing, our approach leads to significant improvements over the state-of-the-art on 27/31 domain shifts from standard UDA benchmarks as well as benchmarks designed to stress-test adaptation under label distribution shift.

(0 min) Generalizing deep neural networks to new target domains is critical to their real-world utility. In practice, it may be feasible to get some target data labeled, but to be cost-effective it is desirable to select a maximally-informative subset via active learning (AL). We study the problem of AL under a domain shift, called Active Domain Adaptation (Active DA). We demonstrate how existing AL approaches based solely on model uncertainty or diversity sampling are less effective for Active DA. We propose Clustering Uncertainty-weighted Embeddings (CLUE), a novel label acquisition strategy for Active DA that performs uncertainty-weighted clustering to identify target instances for labeling that are both uncertain under the model and diverse in feature space. CLUE consistently outperforms competing label acquisition strategies for Active DA and AL across learning settings on 6 diverse domain shifts for image classification.

(0 min) A 3D deep learning model (OARnet) is developed and used to delineate 28 H&N OARs on CT images. OARnet utilizes a densely connected network to detect the OAR bounding-box, then delineates the OAR within the box. It reuses information from any layer to subsequent layers and uses skip connections to combine information from different dense block levels to progressively improve delineation accuracy. Training uses up to 28 expert manual delineated (MD) OARs from 165 CTs. Dice similarity coefficient (DSC) and the 95th percentile Hausdorff distance (HD95) with respect to MD is assessed for 70 other CTs. Mean, maximum, and root-mean-square dose differences with respect to MD are assessed for 56 of the 70 CTs. OARnet is compared with UaNet, AnatomyNet, and Multi-Atlas Segmentation (MAS). Wilcoxon signed-rank tests using 95% confidence intervals are used to assess significance. Wilcoxon signed ranked tests show that, compared with UaNet, OARnet improves (p<0.05) the DSC (23/28 OARs) and HD95 (17/28). OARnet outperforms both AnatomyNet and MAS for DSC (28/28) and HD95 (27/28). Compared with UaNet, OARnet improves median DSC up to 0.05 and HD95 up to 1.5mm. Compared with AnatomyNet and MAS, OARnet improves median (DSC, HD95) by up to (0.08, 2.7mm) and (0.17, 6.3mm). Dosimetrically, OARnet outperforms UaNet (Dmax 7/28; Dmean 10/28), AnatomyNet (Dmax 21/28; Dmean 24/28), and MAS (Dmax 22/28; Dmean 21/28). The DenseNet architecture is optimized using a hybrid approach that performs OAR-specific bounding box detection followed by feature recognition. Compared with other auto-delineation methods, OARnet is better than or equal to UaNet for all but one geometric (Temporal Lobe L, HD95) and one dosimetric (Eye L, mean dose) endpoint for the 28 H&N OARs, and is better than or equal to both AnatomyNet and MAS for all OARs.

(0 min) We present our solutions to the Google Landmark Challenges 2021, for both the retrieval and the recognition tracks. Both solutions are ensembles of transformers and ConvNet models based on Sub-center ArcFace with dynamic margins. Since the two tracks share the same training data, we used the same pipeline and training approach, but with different model selections for the ensemble and different post-processing. The key improvement over last year is newer state-of-the-art vision architectures, especially transformers which significantly outperform ConvNets for the retrieval task. We finished third and fourth places for the retrieval and recognition tracks respectively.

(0 min) Identifying unexpected objects on roads in semantic segmentation (e.g., identifying dogs on roads) is crucial in safety-critical applications. Existing approaches use images of unexpected objects from external datasets or require additional training (e.g., retraining segmentation networks or training an extra network), which necessitate a non-trivial amount of labor intensity or lengthy inference time. One possible alternative is to use prediction scores of a pre-trained network such as the max logits (i.e., maximum values among classes before the final softmax layer) for detecting such objects. However, the distribution of max logits of each predicted class is significantly different from each other, which degrades the performance of identifying unexpected objects in urban-scene segmentation. To address this issue, we propose a simple yet effective approach that standardizes the max logits in order to align the different distributions and reflect the relative meanings of max logits within each predicted class. Moreover, we consider the local regions from two different perspectives based on the intuition that neighboring pixels share similar semantic information. In contrast to previous approaches, our method does not utilize any external datasets or require additional training, which makes our method widely applicable to existing pre-trained segmentation models. Such a straightforward approach achieves a new state-of-the-art performance on the publicly available Fishyscapes Lost & Found leaderboard with a large margin. Our code is publicly available at this $\href{https://github.com/shjung13/Standardized-max-logits}{link}$.

(0 min) Unsupervised Domain Adaptation (UDA) for semantic segmentation has been favorably applied to real-world scenarios in which pixel-level labels are hard to be obtained. In most of the existing UDA methods, all target data are assumed to be introduced simultaneously. Yet, the data are usually presented sequentially in the real world. Moreover, Continual UDA, which deals with more practical scenarios with multiple target domains in the continual learning setting, has not been actively explored. In this light, we propose Continual UDA for semantic segmentation based on a newly designed Expanding Target-specific Memory (ETM) framework. Our novel ETM framework contains Target-specific Memory (TM) for each target domain to alleviate catastrophic forgetting. Furthermore, a proposed Double Hinge Adversarial (DHA) loss leads the network to produce better UDA performance overall. Our design of the TM and training objectives let the semantic segmentation network adapt to the current target domain while preserving the knowledge learned on previous target domains. The model with the proposed framework outperforms other state-of-the-art models in continual learning settings on standard benchmarks such as GTA5, SYNTHIA, CityScapes, IDD, and Cross-City datasets. The source code is available at https://github.com/joonh-kim/ETM.

(0 min) The vast work in Deep Learning (DL) has led to a leap in image denoising research. Most DL solutions for this task have chosen to put their efforts on the denoiser's architecture while maximizing distortion performance. However, distortion driven solutions lead to blurry results with sub-optimal perceptual quality, especially in immoderate noise levels. In this paper we propose a different perspective, aiming to produce sharp and visually pleasing denoised images that are still faithful to their clean sources. Formally, our goal is to achieve high perceptual quality with acceptable distortion. This is attained by a stochastic denoiser that samples from the posterior distribution, trained as a generator in the framework of conditional generative adversarial networks (CGAN). Contrary to distortion-based regularization terms that conflict with perceptual quality, we introduce to the CGAN objective a theoretically founded penalty term that does not force a distortion requirement on individual samples, but rather on their mean. We showcase our proposed method with a novel denoiser architecture that achieves the reformed denoising goal and produces vivid and diverse outcomes in immoderate noise levels.

(0 min) Learning subtle representation about object parts plays a vital role in fine-grained visual recognition (FGVR) field. The vision transformer (ViT) achieves promising results on computer vision due to its attention mechanism. Nonetheless, with the fixed size of patches in ViT, the class token in deep layer focuses on the global receptive field and cannot generate multi-granularity features for FGVR. To capture region attention without box annotations and compensate for ViT shortcomings in FGVR, we propose a novel method named Adaptive attention multi-scale Fusion Transformer (AFTrans). The Selective Attention Collection Module (SACM) in our approach leverages attention weights in ViT and filters them adaptively to correspond with the relative importance of input patches. The multiple scales (global and local) pipeline is supervised by our weights sharing encoder and can be easily trained end-to-end. Comprehensive experiments demonstrate that AFTrans can achieve SOTA performance on three published fine-grained benchmarks: CUB-200-2011, Stanford Dogs and iNat2017.

(0 min) Recent work in equivariant deep learning bears strong similarities to physics. Fields over a base space are fundamental entities in both subjects, as are equivariant maps between these fields. In deep learning, however, these maps are usually defined by convolutions with a kernel, whereas they are partial differential operators (PDOs) in physics. Developing the theory of equivariant PDOs in the context of deep learning could bring these subjects even closer together and lead to a stronger flow of ideas. In this work, we derive a $G$-steerability constraint that completely characterizes when a PDO between feature vector fields is equivariant, for arbitrary symmetry groups $G$. We then fully solve this constraint for several important groups. We use our solutions as equivariant drop-in replacements for convolutional layers and benchmark them in that role. Finally, we develop a framework for equivariant maps based on Schwartz distributions that unifies classical convolutions and differential operators and gives insight about the relation between the two.

(0 min) Deep neural networks (DNNs) have become remarkably successful in data prediction, and have even been used to predict future actions based on limited input. This raises the question: do these systems actually "understand" the event similar to humans? Here, we address this issue using videos taken from an accident situation in a driving simulation. In this situation, drivers had to choose between crashing into a suddenly-appeared obstacle or steering their car off a previously indicated cliff. We compared how well humans and a DNN predicted this decision as a function of time before the event. The DNN outperformed humans for early time-points, but had an equal performance for later time-points. Interestingly, spatio-temporal image manipulations and Grad-CAM visualizations uncovered some expected behavior, but also highlighted potential differences in temporal processing for the DNN.

(0 min) Abdominal ultrasound imaging has been widely used to assist in the diagnosis and treatment of various abdominal organs. In order to shorten the examination time and reduce the cognitive burden on the sonographers, we present a classification method that combines the deep learning techniques and k-Nearest-Neighbor (k-NN) classification to automatically recognize various abdominal organs in the ultrasound images in real time. Fine-tuned deep neural networks are used in combination with PCA dimension reduction to extract high-level features from raw ultrasound images, and a k-NN classifier is employed to predict the abdominal organ in the image. We demonstrate the effectiveness of our method in the task of ultrasound image classification to automatically recognize six abdominal organs. A comprehensive comparison of different configurations is conducted to study the influence of different feature extractors and classifiers on the classification accuracy. Both quantitative and qualitative results show that with minimal training effort, our method can "lazily" recognize the abdominal organs in the ultrasound images in real time with an accuracy of 96.67%. Our implementation code is publicly available at: https://github.com/LeeKeyu/abdominal_ultrasound_classification.

(0 min) In this paper, we develop face.evoLVe -- a comprehensive library that collects and implements a wide range of popular deep learning-based methods for face recognition. First of all, face.evoLVe is composed of key components that cover the full process of face analytics, including face alignment, data processing, various backbones, losses, and alternatives with bags of tricks for improving performance. Later, face.evoLVe supports multi-GPU training on top of different deep learning platforms, such as PyTorch and PaddlePaddle, which facilitates researchers to work on both large-scale datasets with millions of images and low-shot counterparts with limited well-annotated data. More importantly, along with face.evoLVe, images before & after alignment in the common benchmark datasets are released with source codes and trained models provided. All these efforts lower the technical burdens in reproducing the existing methods for comparison, while users of our library could focus on developing advanced approaches more efficiently. Last but not least, face.evoLVe is well designed and vibrantly evolving, so that new face recognition approaches can be easily plugged into our framework. Note that we have used face.evoLVe to participate in a number of face recognition competitions and secured the first place. The version that supports PyTorch is publicly available at https://github.com/ZhaoJ9014/face.evoLVe.PyTorch and the PaddlePaddle version is available at https://github.com/ZhaoJ9014/face.evoLVe.PyTorch/tree/master/paddle. Face.evoLVe has been widely used for face analytics, receiving 2.4K stars and 622 forks.

(0 min) Radiation therapy treatment planning is a complex process, as the target dose prescription and normal tissue sparing are conflicting objectives. Automated and accurate dose prediction for radiation therapy planning is in high demand. In this study, we propose a novel learning-based ensemble approach, named LE-NAS, which integrates neural architecture search (NAS) with knowledge distillation for 3D radiotherapy dose prediction. Specifically, the prediction network first exhaustively searches each block from enormous architecture space. Then, multiple architectures are selected with promising performance and diversity. To reduce the inference time, we adopt the teacher-student paradigm by treating the combination of diverse outputs from multiple searched networks as supervisions to guide the student network training. In addition, we apply adversarial learning to optimize the student network to recover the knowledge in teacher networks. To the best of our knowledge, we are the first to investigate the combination of NAS and knowledge distillation. The proposed method has been evaluated on the public OpenKBP dataset, and experimental results demonstrate the effectiveness of our method and its superior performance to the state-of-the-art method.

(0 min) Interactive facial image manipulation attempts to edit single and multiple face attributes using a photo-realistic face and/or semantic mask as input. In the absence of the photo-realistic image (only sketch/mask available), previous methods only retrieve the original face but ignore the potential of aiding model controllability and diversity in the translation process. This paper proposes a sketch-to-image generation framework called S2FGAN, aiming to improve users' ability to interpret and flexibility of face attribute editing from a simple sketch. The proposed framework modifies the constrained latent space semantics trained on Generative Adversarial Networks (GANs). We employ two latent spaces to control the face appearance and adjust the desired attributes of the generated face. Instead of constraining the translation process by using a reference image, the users can command the model to retouch the generated images by involving the semantic information in the generation process. In this way, our method can manipulate single or multiple face attributes by only specifying attributes to be changed. Extensive experimental results on CelebAMask-HQ dataset empirically shows our superior performance and effectiveness on this task. Our method successfully outperforms state-of-the-art methods on attribute manipulation by exploiting greater control of attribute intensity.

(0 min) This work presented a new drone-based face detection dataset Drone LAMS in order to solve issues of low performance of drone-based face detection in scenarios such as large angles which was a predominant working condition when a drone flies high. The proposed dataset captured images from 261 videos with over 43k annotations and 4.0k images with pitch or yaw angle in the range of -90{\deg} to 90{\deg}. Drone LAMS showed significant improvement over currently available drone-based face detection datasets in terms of detection performance, especially with large pitch and yaw angle. Detailed analysis of how key factors, such as duplication rate, annotation method, etc., impact dataset performance was also provided to facilitate further usage of a drone on face detection.

(0 min) The majority of adversarial machine learning research focuses on additive attacks, which add adversarial perturbation to input data. On the other hand, unlike image recognition problems, only a handful of attack approaches have been explored in the video domain. In this paper, we propose a novel attack method against video recognition models, Multiplicative Adversarial Videos (MultAV), which imposes perturbation on video data by multiplication. MultAV has different noise distributions to the additive counterparts and thus challenges the defense methods tailored to resisting additive adversarial attacks. Moreover, it can be generalized to not only Lp-norm attacks with a new adversary constraint called ratio bound, but also different types of physically realizable attacks. Experimental results show that the model adversarially trained against additive attack is less robust to MultAV.

(2 min) Few-shot learning (FSL) aims to learn a classifier that can be easily adapted to accommodate new tasks not seen during training, given only a few examples. To handle the limited-data problem in few-shot regimes, recent methods tend to collectively use a set of local features to densely represent an image instead of using a mixed global feature. They generally explore a unidirectional query-to-support paradigm in FSL, e.g., find the nearest/optimal support feature for each query feature and aggregate these local matches for a joint classification. In this paper, we propose a new method Mutual Centralized Learning (MCL) to fully affiliate the two disjoint sets of dense features in a bidirectional paradigm. We associate each local feature with a particle that can bidirectionally random walk in a discrete feature space by the affiliations. To estimate the class probability, we propose the features' accessibility that measures the expected number of visits to the support features of that class in a Markov process. We relate our method to learning a centrality on an affiliation network and demonstrate its capability to be plugged in existing methods by highlighting centralized local features. Experiments show that our method achieves the state-of-the-art on both miniImageNet and tieredImageNet.

(2 min) Vision transformers rely on a patch token based self attention mechanism, in contrast to convolutional networks. We investigate fundamental differences between these two families of models, by designing a block sparsity based adversarial token attack. We probe and analyze transformer as well as convolutional models with token attacks of varying patch sizes. We infer that transformer models are more sensitive to token attacks than convolutional models, with ResNets outperforming Transformer models by up to $\sim30\%$ in robust accuracy for single token attacks.

(2 min) This work presents MemX: a biologically-inspired attention-aware eyewear system developed with the goal of pursuing the long-awaited vision of a personalized visual Memex. MemX captures human visual attention on the fly, analyzes the salient visual content, and records moments of personal interest in the form of compact video snippets. Accurate attentive scene detection and analysis on resource-constrained platforms is challenging because these tasks are computation and energy intensive. We propose a new temporal visual attention network that unifies human visual attention tracking and salient visual content analysis. Attention tracking focuses computation-intensive video analysis on salient regions, while video analysis makes human attention detection and tracking more accurate. Using the YouTube-VIS dataset and 30 participants, we experimentally show that MemX significantly improves the attention tracking accuracy over the eye-tracking-alone method, while maintaining high system energy efficiency. We have also conducted 11 in-field pilot studies across a range of daily usage scenarios, which demonstrate the feasibility and potential benefits of MemX.